id: ~deepseek/deepseek-pro-latest name: DeepSeek: DeepSeek Pro Latest -pricing.prompt: 0.00000038808 -pricing.completion: 0.00000116424 +pricing.prompt: 0.0000003865 +pricing.completion: 0.0000029 context_length: 1048576 -top_provider.max_completion_tokens: 393216 +top_provider.max_completion_tokens: 943718
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing and token limits changed materially. Teams budgeting or optimizing for this model need immediate notice. |
| signals | pricing.prompt decreased · pricing.completion increased significantly · max_completion_tokens increased 140% |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: deepseek/deepseek-v4.1-flash name: DeepSeek: DeepSeek V4.1 Flash -pricing.prompt: 0.00000015 -pricing.completion: 0.0000006 +pricing.prompt: 0.00000014 +pricing.completion: 0.00000042 context_length: 1048576 -top_provider.max_completion_tokens: 393216 +top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing dropped on both dimensions; output token limit cut significantly. |
| signals | prompt price decreased · completion price decreased · max_completion_tokens reduced by 66% |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-pro-latest name: DeepSeek: DeepSeek Pro Latest -pricing.prompt: 0.0000004 -pricing.completion: 0.0000043 +pricing.prompt: 0.00000038808 +pricing.completion: 0.00000116424 context_length: 1048576 -top_provider.max_completion_tokens: 384000 +top_provider.max_completion_tokens: 393216
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Pricing per token decreased; max completion tokens increased. Direct impact on cost and capability for builders. |
| signals | pricing.prompt decreased · pricing.completion decreased · max_completion_tokens increased · multiple numeric changes on pricing/limits table |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
name: inclusionAI: Ling 3.0 Flash VL pricing.prompt: 0.00000006 pricing.completion: 0.00000018 -context_length: 131072 +context_length: 262144 top_provider.max_completion_tokens: 32768
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | Context window expansion changes what workloads the model can handle. |
| signals | context_length doubled (131072 → 262144) · model capability change affecting API contract |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: deepseek/deepseek-v4.1-flash name: DeepSeek: DeepSeek V4.1 Flash -pricing.prompt: 0.0000001 -pricing.completion: 0.0000005 +pricing.prompt: 0.00000015 +pricing.completion: 0.0000006 context_length: 1048576 -top_provider.max_completion_tokens: 943718 +top_provider.max_completion_tokens: 393216
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing increased and output capacity reduced. Teams budgeting or selecting models need to know. |
| signals | pricing.prompt increased 50% · pricing.completion increased 20% · max_completion_tokens decreased 58% · three numeric changes affecting cost and capability |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
gpt-realtime-2 2026-05-06 Preview -2026-08-31 — +gpt-realtime-2.1 gpt-realtime-2.1 2026-07-07 @@ Suggest a fix? Last updated on -2026-09-14 +2026-09-21
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New model version added to retirement schedule; prior version's end-of-preview date removed. |
| signals | model version added (gpt-realtime-2.1) · retirement date removed from gpt-realtime-2 · new model row inserted into retirement schedule · date change in footer (2026-09-14 → 2026-09-21) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — September 22, 2026 +https://platform.claude.com/docs/en/release-notes/overview#september-22-2026 +Tue, 22 Sep 2026 00:00:00 GMT + +We've launched Claude Opus 5.5 (claude-opus-5-5), a model for long-running agentic coding and knowledge work. It has a 1M token context window by default, 128k max output tokens, and always-on adaptive thinking, at $4 / $20 USD per MTok (Claude Opus 5 is $5 / $25). Claude Opus 5.5 is available on the Claude API, Claude in Amazon Bedrock, Claude Platform on AWS, Claude on Google Cloud, and Claude in Microsoft Foundry. See What's new in Claude Opus 5.5 for capabilities, API changes, and migration guidance. +On Claude Opus 5.5, thinking can't be disabled: thinking: {"type": "disabled"} and thinking: {"type": "enabled", ...} return a 400 error. Omit the thinking field and control thinking depth with the effort parameter. tool_choice types any and tool also return a 400 error, as on Claude Fable 5.1; use auto with strict tool use. On the Claude API and Google Cloud, computer use on this model requires the computer_toolset_20260801 toolset and the earlier computer_20251124 tool returns a 400 error; on Amazon Bedrock, computer_20251124 keeps working. See the migration guide. +Fast mode (research preview) is available for Claude Opus 5.5 on the Claude API. +Tools can now be defined inside a mid-conversation system message, in beta on the Claude API with the inline-tools-2026-09-15 beta header. A tool_addition block can carry the tool's full definition (tool: {"type": "tool_definition", "definition": {...}}), so you can add a tool, change its schema, or move a server tool to a newer version without editing tools or invalidating the prompt cache. The same header covers adding and removing tools by reference. With the MCP connector's mcp-client-2026-09-15 beta header as well, the definition can be an MCP toolset, and a response records each server's fetched tool list in an mcp_tool_listing block, which pins that list when you send it back.
| triage verdict | MATERIAL at 0.98, needed 0.45 |
|---|---|
| triage reasoning | New model release with pricing, breaking API changes, and feature availability across platforms. |
| signals | new model released (Claude Opus 5.5) · pricing change (Opus 5.5: $4/$20 vs Opus 5: $5/$25) · breaking API changes (thinking parameter disabled, tool_choice restrictions) · context window specification (1M tokens) · max output tokens specified (128k) · new feature availability (fast mode, inline tools beta) · toolset deprecation (computer_20251124 returns 400 on Claude API) · multi-platform availability statement |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: z-ai/glm-5.3:batch name: Z.ai: GLM 5.3 (batch) -pricing.prompt: 0.00000072 -pricing.completion: 0.0000024 +pricing.prompt: 0.00000045 +pricing.completion: 0.000002 context_length: 1048576 top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing changes on a batch model directly impact engineering budget and cost calculations. |
| signals | pricing.prompt decreased 37.5% (0.00000072 → 0.00000045) · pricing.completion decreased 16.7% (0.0000024 → 0.000002) · batch model pricing change affects cost planning |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-v4-flash-latest name: DeepSeek: DeepSeek V4 Flash Latest -pricing.prompt: 0.00000003 -pricing.completion: 0.000001 +pricing.prompt: 0.000000038 +pricing.completion: 0.00000055 context_length: 1310720 top_provider.max_completion_tokens: 943718
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing changes directly affect cost calculations and budget planning for teams using this model. |
| signals | pricing.prompt decreased ~21% · pricing.completion decreased ~45% · both token prices moved in same direction (down) · numerical precision increased (more decimal places) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: nvidia/nemotron-3.5-lightning name: NVIDIA: Nemotron 3.5 Lightning -pricing.prompt: 0.00000007 +pricing.prompt: 0.00000008 pricing.completion: 0.0000002 context_length: 262144 -top_provider.max_completion_tokens: 235929 +top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Pricing increase and substantial reduction in max completion tokens affect cost and capability planning. |
| signals | prompt pricing increased · max_completion_tokens decreased significantly (235929 → 131072, ~44% reduction) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
pricing.prompt: 0.000000075 pricing.completion: 0.00000025 context_length: 1310720 -top_provider.max_completion_tokens: 943718 +top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Model's maximum completion token output reduced by 86%, directly impacts application design and token budgeting. |
| signals | max_completion_tokens decreased significantly · 943718 → 131072 is ~7x reduction · affects model capability / output limits |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-pro-latest name: DeepSeek: DeepSeek Pro Latest -pricing.prompt: 0.00000039996 -pricing.completion: 0.00000119988 +pricing.prompt: 0.0000004 +pricing.completion: 0.0000043 context_length: 1048576 -top_provider.max_completion_tokens: 393216 +top_provider.max_completion_tokens: 384000
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing increased significantly and output token limit reduced. Teams budgeting or optimizing for this model need to recalculate. |
| signals | pricing.prompt increased · pricing.completion increased ~3.6x · max_completion_tokens decreased by 9216 tokens |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: deepseek/deepseek-v4.1-flash name: DeepSeek: DeepSeek V4.1 Flash -pricing.prompt: 0.00000015 -pricing.completion: 0.0000006 +pricing.prompt: 0.0000001 +pricing.completion: 0.0000005 context_length: 1048576 -top_provider.max_completion_tokens: 384000 +top_provider.max_completion_tokens: 943718
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing and token limits changed materially for an active model. |
| signals | prompt price decreased 33% · completion price decreased 17% · max_completion_tokens increased 146% |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-pro-latest name: DeepSeek: DeepSeek Pro Latest -pricing.prompt: 0.00000056364 -pricing.completion: 0.00000169092 +pricing.prompt: 0.00000039996 +pricing.completion: 0.00000119988 context_length: 1048576 -top_provider.max_completion_tokens: 384000 +top_provider.max_completion_tokens: 393216
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing dropped significantly and token limits increased. Teams budgeting or optimizing costs need to know. |
| signals | prompt price decreased 29% · completion price decreased 29% · max_completion_tokens increased from 384000 to 393216 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~anthropic/claude-opus-latest name: Anthropic: Claude Opus Latest -pricing.prompt: 0.000005 -pricing.completion: 0.000025 +pricing.prompt: 0.000004 +pricing.completion: 0.00002 context_length: 1000000 top_provider.max_completion_tokens: 128000
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Pricing for Claude Opus Latest decreased. Teams budgeting or optimizing costs need to know. |
| signals | price decreased on prompt tokens (0.000005 → 0.000004) · price decreased on completion tokens (0.000025 → 0.00002) · pricing change affects cost calculations for active deployments |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+id: openai/gpt-6-sol:batch +name: OpenAI: GPT-6 Sol (batch) +pricing.prompt: 0.000001 +pricing.completion: 0.000005 +context_length: 1050000 +top_provider.max_completion_tokens: 128000
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New model entry with pricing and limits affects platform availability and cost planning for engineers. |
| signals | new model added to catalogue · pricing specified (prompt and completion rates) · context length declared (1.05M tokens) · completion token limit specified (128k) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+id: openai/gpt-6-sol-pro:batch +name: OpenAI: GPT-6 Sol Pro (batch) +pricing.prompt: 0.000001 +pricing.completion: 0.000005 +context_length: 1050000 +top_provider.max_completion_tokens: 128000
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New model entry with pricing and limits affects platform selection and cost planning. |
| signals | new model added to catalogue · pricing specified (prompt and completion rates) · context length defined · completion token limit specified · batch variant indicates new capability tier |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+id: openai/gpt-6-sol-pro +name: OpenAI: GPT-6 Sol Pro +pricing.prompt: 0.000002 +pricing.completion: 0.00001 +context_length: 1050000 +top_provider.max_completion_tokens: 128000
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New model entry with pricing and limits affects platform availability and cost planning. |
| signals | new model added to catalogue · pricing specified (prompt and completion rates) · context length defined · completion token limit set |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+id: openai/gpt-6-luna:batch +name: OpenAI: GPT-6 Luna (batch) +pricing.prompt: 0.00000005 +pricing.completion: 0.00000025 +context_length: 1050000 +top_provider.max_completion_tokens: 128000
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New model entry with pricing and limits affects platform availability and cost planning for engineers. |
| signals | new model added to catalogue · pricing specified (prompt and completion rates) · context length defined · completion token limit specified |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+id: anthropic/claude-opus-5.5:batch +name: Anthropic: Claude Opus 5.5 (batch) +pricing.prompt: 0.000002 +pricing.completion: 0.00001 +context_length: 1000000 +top_provider.max_completion_tokens: 128000
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | New model availability with defined pricing and limits affects platform selection and cost planning. |
| signals | new model added to catalogue · pricing specified (prompt and completion rates) · context window defined (1M tokens) · completion token limit specified (128k) · batch variant of existing model |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~z-ai/glm-latest name: Z.ai: GLM Latest -pricing.prompt: 0.0000006545 -pricing.completion: 0.000002057 +pricing.prompt: 0.0000006538 +pricing.completion: 0.0000020548 context_length: 1310720 -top_provider.max_completion_tokens: 943718 +top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing dropped and max output tokens cut dramatically. Teams budgeting or relying on long completions need to know. |
| signals | pricing.prompt decreased · pricing.completion decreased · max_completion_tokens reduced by 86% (943718 → 131072) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-pro-latest name: DeepSeek: DeepSeek Pro Latest -pricing.prompt: 0.000000624 -pricing.completion: 0.00000288 +pricing.prompt: 0.00000056364 +pricing.completion: 0.00000169092 context_length: 1048576 -top_provider.max_completion_tokens: 943718 +top_provider.max_completion_tokens: 384000
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing cuts and token limit reduction directly impact engineering budgets and model capability planning. |
| signals | prompt pricing decreased ~9.8% · completion pricing decreased ~41.3% · max_completion_tokens reduced from 943718 to 384000 (~59% reduction) · three numeric changes affecting cost and capability |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: z-ai/glm-5.3:batch name: Z.ai: GLM 5.3 (batch) -pricing.prompt: 0.0000007 -pricing.completion: 0.0000022 +pricing.prompt: 0.00000072 +pricing.completion: 0.0000024 context_length: 1048576 -top_provider.max_completion_tokens: 943718 +top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Pricing increased and max output tokens cut dramatically, affecting cost and feasibility of workloads. |
| signals | pricing.prompt increased 2.9% · pricing.completion increased 9.1% · max_completion_tokens decreased 86% (943718 → 131072) · breaking change to output capacity |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: z-ai/glm-5.3-flash:batch name: Z.ai: GLM 5.3 Flash (batch) -pricing.prompt: 0.000000075 -pricing.completion: 0.00000025 +pricing.prompt: 0.00000006 +pricing.completion: 0.0000002 context_length: 1048576 -top_provider.max_completion_tokens: 943718 +top_provider.max_completion_tokens: 131072
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing dropped and max output tokens cut dramatically. Teams budgeting or relying on long completions must know. |
| signals | pricing.prompt decreased 20% · pricing.completion decreased 20% · max_completion_tokens reduced by 86% (943718 → 131072) · breaking change to model capability |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: qwen/qwen3.6-27b name: Qwen: Qwen3.6 27B -pricing.prompt: 0.0000003 -pricing.completion: 0.000002 +pricing.prompt: 0.00000032 +pricing.completion: 0.0000027 context_length: 262144 -top_provider.max_completion_tokens: 65536 +top_provider.max_completion_tokens: 262140
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing increased and output token limit quadrupled. Direct impact on engineering budgets and application design. |
| signals | prompt pricing increased ~6.7% · completion pricing increased ~35% · max_completion_tokens increased 4x (65536 → 262140) · pricing and limits are explicit commitments affecting cost and capability |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: openai/gpt-oss-20b name: OpenAI: gpt-oss-20b -pricing.prompt: 0.00000003 -pricing.completion: 0.00000013 +pricing.prompt: 0.000000018 +pricing.completion: 0.00000009 context_length: 131072 -top_provider.max_completion_tokens: 117964 +top_provider.max_completion_tokens: 32768
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing dropped significantly and max output tokens cut by 72%, affecting budget and use-case viability. |
| signals | pricing.prompt decreased 40% · pricing.completion decreased 31% · max_completion_tokens reduced by 72% (117964 → 32768) · model capability constraint change |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
name: AionLabs: Aion-3.0-Mini pricing.prompt: 0.0000007 pricing.completion: 0.0000014 -context_length: 131072 +context_length: 1048576 top_provider.max_completion_tokens: 32768
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Context window expansion is a material capability change affecting what workloads the model can handle. |
| signals | context_length increased 8x (131072 → 1048576 tokens) · model capability change affects application architecture decisions |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
name: AionLabs: Aion-3.0 pricing.prompt: 0.000003 pricing.completion: 0.000006 -context_length: 131072 +context_length: 1048576 top_provider.max_completion_tokens: 32768
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Context window expansion is a material capability change affecting prompt design and use-case feasibility. |
| signals | context_length increased 8x (131072 → 1048576 tokens) · model capability change affects application design decisions |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
name: AionLabs: Aion-2.0 pricing.prompt: 0.0000008 pricing.completion: 0.0000016 -context_length: 131072 +context_length: 1048576 top_provider.max_completion_tokens: 32768
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Context window expansion is a material capability change affecting how engineers can use this model. |
| signals | context_length increased 8x (131072 → 1048576 tokens) · model capability change affects application design decisions |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
pricing.prompt: 0.000000075 pricing.completion: 0.00000025 context_length: 1310720 -top_provider.max_completion_tokens: 102400 +top_provider.max_completion_tokens: 943718
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | Completion token limit nearly 10x higher; engineering teams must re-evaluate request sizing and costs. |
| signals | max_completion_tokens increased 9.2x (102400 → 943718) · limit change on production model · affects token budgeting and request planning |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-v4-flash-latest name: DeepSeek: DeepSeek V4 Flash Latest pricing.prompt: 0.00000003 -pricing.completion: 0.0000008 +pricing.completion: 0.000001 context_length: 1310720 top_provider.max_completion_tokens: 943718
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Completion pricing for DeepSeek V4 Flash increased 25%, affecting cost calculations for production workloads. |
| signals | completion token price increased · price change: 0.0000008 → 0.000001 (25% increase) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
id: ~deepseek/deepseek-pro-latest name: DeepSeek: DeepSeek Pro Latest -pricing.prompt: 0.000000558624 -pricing.completion: 0.000001675872 +pricing.prompt: 0.000000624 +pricing.completion: 0.00000288 context_length: 1048576 -top_provider.max_completion_tokens: 384000 +top_provider.max_completion_tokens: 943718
| triage verdict | MATERIAL at 0.98, needed 0.5 |
|---|---|
| triage reasoning | Pricing per token increased significantly; max output tokens nearly doubled. |
| signals | pricing.prompt increased ~11.7% · pricing.completion increased ~71.6% · max_completion_tokens increased ~145% (384000 → 943718) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
pricing.prompt: 0.00000015 pricing.completion: 0.0000012 context_length: 262144 -top_provider.max_completion_tokens: 32768 +top_provider.max_completion_tokens: 235929
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | Maximum completion token limit changed substantially, affecting application design and output capacity. |
| signals | max_completion_tokens decreased significantly · 32768 → 235929 represents constraint change · affects output capacity for dependent applications |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
https://platform.claude.com/docs/en/release-notes/overview#september-14-2026 Mon, 14 Sep 2026 00:00:00 GMT -The Messages API can now compact a conversation on demand on the Claude API, in beta with the compact-2026-09-04 beta header. Send the top-level compaction parameter, and the API returns a signed compaction block that summarizes the messages you sent. On later requests, send that block first, in place of those messages. You choose when to compact, the request can run in the background, and you can keep recent turns word for word after the summary. On models with preserved thinking, the thinking in those kept turns can stay valid. +The Messages API can now compact a conversation on demand on the Claude API, in beta with the compact-2026-09-04 beta header. Send the top-level compaction parameter, and the API returns a signed compaction block that summarizes the messages you sent. On later requests, send that block first, in place of those messages. You choose when to compact, the request can run in the background, and you can keep recent turns word for word after the summary. On models with preserved thinking, the thinking in those kept turns can stay valid. +With the thinking-binding-controls-2026-08-01 beta header, the input_transformations response field gains a second entry type, thinking_mismatch_allowed. It names a thinking block that failed the prefix check on a request where the API doesn't enforce that check: on Claude Fable 5.1, for example, a request from an account created before August 31, 2026, with prefix_mismatch_behavior unset. The block still reaches the model unchanged. Log these entries to find history edits in production traffic before you opt into enforcement. See Set the mismatch behavior and read input_transformations.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New beta API feature with date-gated behavior and enforcement controls affects how engineers handle thinking blocks. |
| signals | new beta feature with specific header (thinking-binding-controls-2026-08-01) · new response field type (thinking_mismatch_allowed) · new API behavior affecting model input handling · date-gated behavior (accounts created before August 31, 2026) · enforcement opt-in mechanism introduced · breaking change potential for existing thinking implementations |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
-IndividualDeveloperTeam & EnterpriseAPI +IndividualTeam & EnterpriseAPI -Individual - Free Try Claude @@ Create Artifacts Pro -Claude Code included For everyday work $17 @@ More usage* Hand off and schedule tasks Claude Design, Slides, Docs +Claude Code Claude Science Projects More Claude models Claude in Chrome and Microsoft 365 Max -Claude Code included For people who work with Claude all day From $100
| triage verdict | MATERIAL at 0.92, needed 0.55 |
|---|---|
| triage reasoning | Pricing tier restructuring and feature redistribution across tiers affects plan selection and cost. |
| signals | pricing tier renamed: 'Individual' removed, 'Team' introduced · feature moved between tiers: 'Claude Code' removed from Pro description, added to Max description · tier structure simplified: 'IndividualDeveloperTeam' collapsed to 'Team' · feature availability changed: Claude Code no longer included in Pro tier |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
gpt-realtime-2.1 2026-07-07 -Preview +GA 2027-06-25 — gpt-realtime-2.1-mini 2026-07-07 -Preview +GA 2027-06-25 —
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | Preview-to-GA transition signals production-ready models with different support commitments and retirement timelines. |
| signals | model status change from Preview to GA · two models affected · GA status affects support SLA and production readiness |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
-IndividualDeveloperTeam & EnterpriseAPI +IndividualTeam & EnterpriseAPI -Individual - Free Try Claude @@ Create Artifacts Pro +Claude Code included For everyday work $17 @@ More usage* Hand off and schedule tasks Claude Design, Slides, Docs -Claude Code Claude Science Projects More Claude models Claude in Chrome and Microsoft 365 Max +Claude Code included For people who work with Claude all day From $100
| triage verdict | MATERIAL at 0.92, needed 0.55 |
|---|---|
| triage reasoning | Claude Code inclusion moved between pricing tiers, altering what features are available at each price point. |
| signals | pricing tier restructuring · feature moved between tiers · Claude Code added to Pro tier · Claude Code removed from Max tier description · tier naming changed (Individual → removed, Developer removed) · feature availability changed across product lines |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
gpt-4.1-nano 2025-04-14 Legacy -2026-10-14 +2027-04-14 — gpt-4o
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Retirement deadline for gpt-4.1-nano moved from Oct 2026 to Apr 2027, affecting deployment planning. |
| signals | retirement date changed · sunset date extended by 6 months · model lifecycle impact |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
-IndividualDeveloperTeam & EnterpriseAPI +IndividualTeam & EnterpriseAPI -Individual - Free Try Claude @@ Free for everyone Try Claude -Chat on web, iOS, Android, and on your desktop -Generate code and visualize data -Write, edit, and create content -Ability to search the web +Chat on web, desktop, and mobile +Search the web, create files, and run code Memory across conversations -Create files and execute code -Unlock more from Claude with desktop extensions -Connect Slack and Google Workspace services -Integrate any context or tool through connectors with remote MCP -Extended thinking for complex work +Connect your apps and tools +Create Artifacts Pro -For everyday productivity +For everyday work $17 Per month with annual subscription discount ($200 billed up front). $20 if billed monthly. @@ Try Claude Everything in Free, plus: More usage* -Includes Claude Code -Includes Claude Cowork -Includes Claude Design -Includes Claude Science -Access to unlimited projects to organize chats and documents -Access to Research -Ability to use more Claude models -Claude for Microsoft 365 +Hand off and schedule tasks +Claude Design, Slides, Docs +Claude Code +Claude Science +Projects +More Claude models +Claude in Chrome and Microsoft 365 Max -Get the most out of Claude +For people who work with Claude all day From $100 Per month @@ Features and capabilities FeaturesFreeProMax 5xMax 20x -Chat on web, iOS, Android and Desktop +@Claude +No +No +No +No +Claude Code + +No Yes Yes Yes -Yes -Claude Code +Claude Science No Yes Yes Yes -Claude CoworkMore information +Claude in Chrome No Yes Yes Yes -Claude Design +Claude for Microsoft 365 No Yes Yes Yes -Claude Science +Claude for Microsoft Outlook No Yes Yes Yes -@Claude +Chat on web, iOS, Android and Desktop +Yes +Yes +Yes +Yes +Claude Design, Slides, Docs + No -No -No -No +Yes +Yes +Yes Ability to search the web Yes Yes Yes Yes -Desktop extensions +Create ArtifactsMore information Yes Yes Yes Yes -Voice mode +Desktop extensions Yes Yes Yes Yes -Incognito chatsMore information +Voice mode Yes Yes Yes Yes -User preferences +Incognito chatsMore information Yes Yes Yes Yes -ArtifactsMore information +User preferences Yes Yes Yes Yes Projects - -No +Up to 5 Yes Yes Yes @@ No No No -Claude in Chrome -No -Yes -Yes -Yes -Claude for Microsoft 365 - -No -Yes -Yes -Yes -Claude for Microsoft Outlook - -No -Yes -Yes -Yes - Security and administration FeaturesFreeProMax 5xMax 20x Single sign-on (SSO) and domain capture @@ No No Billing cycle -n/aMonthly and annualMonthly and annualMonthly and annual +n/aMonthly and annualMonthlyMonthly Adding seats midterm n/an/an/an/a @@ Yes Yes Context windowMore information -200k200k200k200k +Up to 1Mvaries by modelUp to 1Mvaries by modelUp to 1Mvaries by modelUp to 1Mvaries by model Usage creditsMore information No
| triage verdict | MATERIAL at 0.92, needed 0.55 |
|---|---|
| triage reasoning | Pricing tier restructuring, feature removals, context window expansion, and billing cycle restrictions materially affect platform capabilities and costs. |
| signals | pricing tier renamed: 'Individual Developer' removed, 'Team' tier introduced · feature set reorganized with removals and additions across tiers · context window changed from fixed '200k' to 'Up to 1M varies by model' · billing cycle for Max tiers changed from 'Monthly and annual' to 'Monthly only' · projects limit changed from unlimited to 'Up to 5' in Free tier · Claude Cowork feature removed entirely · new features added: Claude Design/Slides/Docs, Claude in Chrome, Claude for Outlook |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
gpt-4.1-nano 2025-04-14 Legacy -2027-04-14 +2026-10-14 — gpt-4o @@ gpt-image-1.5 2025-12-16 GA -2027-06-16 +2026-12-16 — gpt-image-2 @@ Cohere-command-a-plus-05-2026 1 Preview -2026-10-13 -— +2026-10-16 +coherelabs-command-a-plus-05-2026-fp81, coherelabs-command-a-plus-05-2026-bf161, coherelabs-command-a-plus-05-2026-w4a41 embed-v-4-0 1 @@ — — +1 Available through the Hugging Face collection, for managed compute deployment. + DeepSeek Model @@ Suggest a fix? Last updated on -2026-09-02 +2026-09-14
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Two model retirement dates moved forward by 6 months each, requiring engineering action. |
| signals | retirement date accelerated: gpt-4.1-nano from 2027-04-14 to 2026-10-14 (6 months earlier) · retirement date accelerated: gpt-image-1.5 from 2027-06-16 to 2026-12-16 (6 months earlier) · preview model date changed: Cohere-command-a-plus-05-2026 from 2026-10-13 to 2026-10-16 · model variant identifiers added for Cohere model · new deployment option noted for embed-v-4-0 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — September 14, 2026 +https://platform.claude.com/docs/en/release-notes/overview#september-14-2026 +Mon, 14 Sep 2026 00:00:00 GMT + +The Messages API can now compact a conversation on demand on the Claude API, in beta with the compact-2026-09-04 beta header. Send the top-level compaction parameter, and the API returns a signed compaction block that summarizes the messages you sent. On later requests, send that block first, in place of those messages. You choose when to compact, the request can run in the background, and you can keep recent turns word for word after the summary. On models with preserved thinking, the thinking in those kept turns can stay valid.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New API capability with beta header and parameter changes how developers structure conversation requests. |
| signals | new API feature announced · beta header introduced (compact-2026-09-04) · new top-level parameter (compaction) · changes conversation handling workflow · affects request structure and API contract |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Upcoming deprecations are listed below, with the most recent announcements at the top. +2026-09-11: GPT-5.4-Cyber + +The gpt-5.4-cyber model is deprecated and will be removed from the API on October 1, 2026. Migrate to gpt-5.6-cyber before the shutdown date. + +Shutdown dateModel / systemRecommended replacement +Oct 1, 2026gpt-5.4-cybergpt-5.6-cyber + 2026-08-26: Transcription models On August 26, 2026, we notified developers using whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe, and gpt-4o-transcribe-diarize of their deprecation and removal from the API on February 26, 2027. @@ The Realtime API Beta was deprecated and removed from the API on May 12, 2026. -There are a few key differences between the interfaces in the Realtime beta API and the released GA API. See the migration guide for the current GA interface and related Realtime docs. +The interfaces in the Realtime beta API and the released GA API have a few key differences. See the migration guide for the current GA interface and related Realtime docs. Shutdown dateModel / systemRecommended replacement 2026‐05‐12OpenAI-Beta: realtime=v1Realtime API
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | New deprecation entry for gpt-5.4-cyber with October 2026 shutdown date requires migration action. |
| signals | new model deprecation announced · future shutdown date (Oct 1, 2026) · migration path specified (gpt-5.4-cyber → gpt-5.6-cyber) · model removal from API · sentence reword is cosmetic noise |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
o1 2024-12-17 Deprecated -2026-10-21 +2026-11-19 gpt-5.6-sol o1-pro 2025-03-19 -GA -2026-10-21 +Deprecated +2026-11-19 gpt-5.6-sol o3 2025-04-16 -GA -2026-10-21 +Deprecated +2026-11-19 gpt-5.6-sol o3-deep-research 2025-06-26 -GA -2026-12-26 -— +Deprecated +2026-11-19 +gpt-5.6-sol o3-mini 2025-01-31 Deprecated -2026-10-01 -o4-mini +2026-11-19 +gpt-5.6-terra o3-pro 2025-06-10 -GA -2026-12-17 -— +Deprecated +2026-11-19 +gpt-5.6-sol o4-mini 2025-04-16 Deprecated -2026-10-16 -— +2026-11-19 +gpt-5.6-terra sora-2 2025-10-06 @@ o4-mini 2025-04-16 -No earlier than 2027-04-161 -2027-10-16 +No earlier than 2027-05-191 +2027-11-19 1 For existing customers only. Otherwise, training retirement occurs at base model retirement.
| triage verdict | MATERIAL at 0.98, needed 0.45 |
|---|---|
| triage reasoning | Multiple model deprecations accelerated and status changes from GA to Deprecated affect deployment planning. |
| signals | multiple model deprecation dates shifted forward by ~30 days · models changed from GA to Deprecated status · o1-pro, o3, o3-deep-research, o3-pro transitioned from GA to Deprecated · successor model assignments changed (o3-mini now maps to gpt-5.6-terra instead of o4-mini) · training retirement dates shifted (2027-04-16 → 2027-05-19, 2027-10-16 → 2027-11-19) · consolidation: multiple models now share single sunset date 2026-11-19 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — September 10, 2026 +https://platform.claude.com/docs/en/release-notes/overview#september-10-2026 +Thu, 10 Sep 2026 00:00:00 GMT + +Claude Managed Agents permission policies now include auto: the server evaluates each agent or MCP tool call and runs it, denies it, or pauses for your approval. agent.tool_use and agent.mcp_tool_use events report how each call was evaluated in an evaluation field alongside evaluated_permission. See Let the server evaluate each call with auto. +The ant CLI adds ant beta:sessions connect, which attaches your terminal to a Claude Managed Agents session. You can follow the session live, send messages, and allow or deny tool calls that are waiting for approval. Pass --web to serve the Claude Console's session viewer locally and open the session there instead. See Connect to a Managed Agents session from your terminal.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New agent permission modes, API fields, and CLI commands affect how teams integrate and control Claude Managed Agents. |
| signals | new feature: auto permission policy mode for Claude Managed Agents · new API field: evaluation field in agent.tool_use and agent.mcp_tool_use events · new CLI command: ant beta:sessions connect · new capability: terminal attachment to live agent sessions · new parameter: --web flag for local console serving |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
2027-10-21 — +gpt-image-2.5-flare +2026-09-09 +GA +2027-09-09 +— + +gpt-image-2.5-sunburst +2026-09-09 +GA +2027-09-09 +— + gpt-realtime 2025-08-28 GA
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | New models with defined sunset dates added to retirement schedule; engineering teams must track these timelines. |
| signals | two new models added to retirement schedule · future sunset dates specified (2027-09-09) · GA status indicates production-ready models · retirement schedule is the core tracked artifact |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
https://platform.claude.com/docs/en/release-notes/overview#september-3-2026 Thu, 03 Sep 2026 00:00:00 GMT -Version 1.30.0 of the ant CLI adds ant apply, which creates and updates agents, environments, skills, memory stores, and deployments from files in your repository. Describe each resource in a file, run ant apply, and approve the plan it prints. Commit the claude-lock.json lockfile it writes so that later runs, on your machine or in CI, update the same resources instead of creating new ones. See Manage resources as code with ant apply. +Version 1.30.0 of the ant CLI adds ant apply, which creates and updates agents, environments, skills, memory stores, and deployments from files in your repository. Describe each resource in a file, run ant apply, and approve the plan it prints. Commit the claude-lock.json lockfile it writes so that later runs, on your machine or in CI, update the same resources instead of creating new ones. See Manage resources as code with ant apply. +Per-message effort changes, in beta, are also available on Google Cloud for Claude Fable 5.1, Claude Mythos 5.1, and Claude Opus 5, with the same mid-conversation-output-config-2026-07-01 beta header.
| triage verdict | MATERIAL at 0.72, needed 0.45 |
|---|---|
| triage reasoning | New beta feature (per-message effort changes) now available on Google Cloud for specific Claude models with required API header. |
| signals | new feature announcement · beta feature availability on new platform · new models listed with feature support · specific API header requirement mentioned |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Includes Claude Design Includes Claude Science Access to unlimited projects to organize chats and documents -Access to Research Ability to use more Claude models Claude for Microsoft 365 @@ Yes Yes Yes -Research - -No -Yes -Yes -Yes Memory Yes
| triage verdict | MATERIAL at 0.85, needed 0.55 |
|---|---|
| triage reasoning | Research feature access removed from a pricing tier, affecting what customers can do. |
| signals | feature removed from pricing tier · Research access removed · tier comparison table altered |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
-Pricing +IndividualTeam & EnterpriseAPI -Individual - -Team & Enterprise - -API - -Active - Free Try Claude $0 Free for everyone -Try ClaudeTry Claude - +Try Claude Chat on web, iOS, Android, and on your desktop Generate code and visualize data Write, edit, and create content @@ $17 Per month with annual subscription discount ($200 billed up front). $20 if billed monthly. -Try ClaudeTry Claude - +Try Claude Everything in Free, plus: More usage* Includes Claude Code @@ From $100 Per month -Try ClaudeTry Claude - +Try Claude Everything in Pro, plus: Choose 5x or 20x more usage than Pro* Higher output limits for all tasks @@ Priority access at high traffic times Usage limits apply. Prices shown don’t include applicable tax. Price and plans are subject to change at Anthropic's discretion. - Usage limits apply. Price and plans are subject to change at Anthropic's discretion. -Compare features across plans - Search Free +Try Claude -Try ClaudeTry Claude - Pro +Try Claude -Try ClaudeTry Claude - Max 5x +Try Claude -Try ClaudeTry Claude - Max 20x +Try Claude -Try ClaudeTry Claude - Features and capabilities - FeaturesFreeProMax 5xMax 20x Chat on web, iOS, Android and Desktop +Yes +Yes +Yes +Yes Claude Code -Claude Cowork +No +Yes +Yes +Yes +Claude CoworkMore information +No +Yes +Yes +Yes Claude Design +No +Yes +Yes +Yes Claude Science +No +Yes +Yes +Yes @Claude +No +No +No +No Ability to search the web +Yes +Yes +Yes +Yes Desktop extensions +Yes +Yes +Yes +Yes Voice mode -Incognito chats +Yes +Yes +Yes +Yes +Incognito chatsMore information +Yes +Yes +Yes +Yes User preferences -Artifacts +Yes +Yes +Yes +Yes +ArtifactsMore information +Yes +Yes +Yes +Yes Projects +No +Yes +Yes +Yes Project sharing and collaboration -Create and edit files with code execution +No +No +No +No +Create and edit files with code executionMore information +Yes +Yes +Yes +Yes Research +No +Yes +Yes +Yes Memory -Skills +Yes +Yes +Yes +Yes +SkillsMore information -Connectors +Yes +Yes +Yes +Yes +ConnectorsMore information +Yes +Yes +Yes +Yes Web search -Enterprise search +Yes +Yes +Yes +Yes +Enterprise searchMore information +No +No +No +No Claude in Chrome +No +Yes +Yes +Yes Claude for Microsoft 365 +No +Yes +Yes +Yes Claude for Microsoft Outlook -Security and administration +No +Yes +Yes +Yes +Security and administration FeaturesFreeProMax 5xMax 20x Single sign-on (SSO) and domain capture +No +No +No +No Central billing and administration +No +No +No +No Admin controls for remote and local connectors +No +No +No +No Enterprise deployment for the Claude desktop app +No +No +No +No Role-based access +No +No +No +No System for Cross-domain Identity Management (SCIM) +No +No +No +No Audit logs +No +No +No +No Usage analytics +No +No +No +No Compliance API +No +No +No +No HIPAA-ready offering +No +No +No +No Custom data retention controls +No +No +No +No Organization wide skills deployment -Payment options +No +No +No +No +Payment options FeaturesFreeProMax 5xMax 20x Credit card n/a - +Yes +Yes +Yes ACH +No +No +No +No Invoicing / net terms +No +No +No +No Billing cycle -n/a - -Monthly and annual - -Monthly and annual - -Monthly and annual - +n/aMonthly and annualMonthly and annualMonthly and annual Adding seats midterm -n/a +n/an/an/an/a -n/a - -n/a - -n/a - Models and usage - FeaturesFreeProMax 5xMax 20x -Fable +FableMore information -Usage credits - -50% of weekly limits* - -50% of weekly limits* - +NoUsage credits50% of weekly limits*50% of weekly limits* Opus +No +Yes +Yes +Yes Sonnet +Yes +Yes +Yes +Yes Haiku -Context window +Yes +Yes +Yes +Yes +Context windowMore information +200k200k200k200k +Usage creditsMore information -200k - -200k - -200k - -200k - -Usage credits - +No +Yes +Yes +Yes Priority access at high traffic times +No +No +Yes +Yes User and organizational level spend controls -Model training +No +No +No +No +Model trainingMore information +Opt-outOpt-outOpt-outOpt-out -Opt-out +Plans and usageBilling and paymentsManaging your plan -Opt-out - -Opt-out - -Opt-out - -Team -For teams of 2 to 150 - -Get Team planGet Team plan - -Standard seat -All Claude features, plus more usage than Pro* - -$20 - -Per seat / month if billed annually. $25 if billed monthly. - -Premium seat -5x more usage than standard seats* - -$100 - -Per seat / month if billed annually. $125 if billed monthly. - -Includes Claude Code and Claude Cowork -Includes Claude Design -Includes Claude Science -Connect Microsoft 365, @Claude, and more -Enterprise search across your organization -Central billing and administration -Single sign-on (SSO) -Admin controls for remote and local connectors -Enterprise deployment for the Claude desktop app -No model training on your content by default -Mix and match seat types - -Enterprise -For large businesses operating at scale - -Get Enterprise planGet Enterprise plan - -Chat with buying specialistChat with buying specialist - -Seat price + usage at API rates - -$20/seat. Usage cost scales with model and task. - -All Team plan features, plus: -Admins set user and org spend limits -Role-based access with fine grained permissioning -System for Cross-domain Identity Management (SCIM) -Audit logs -Compliance API for observability and monitoring -Custom data retention controls -Network-level access control -IP allowlisting -HIPAA-ready offering available -Claude Security (beta) - -Usage limits apply. Prices shown don’t include applicable tax. Price and plans are subject to change at Anthropic's discretion. - -Usage limits apply. Price and plans are subject to change at Anthropic's discretion. - -Education plan - -Get a comprehensive university‐wide plan for an institution, including its students, faculty, and staff. - -Learn moreLearn more - -Student and faculty access - -Comprehensive access for all university members at discounted rates - -Academic research and learning mode - -Dedicated API credits and educational features for student learning - -Training and enablement - -Resources for successful adoption across your institution - -Compare features across plans - -Search - -Team - -Create planCreate plan - -Enterprise (self-serve) - -Create planCreate plan - -Enterprise (sales-assisted) - -Contact salesContact sales - -Features and capabilities - -FeaturesTeamEnterprise (self-serve)Enterprise (sales-assisted) -Chat on web, iOS, Android and Desktop - -Claude Code - -Claude Cowork - -Claude Design - -Claude Science - -@Claude - -Claude Security - -Ability to search the web - -Desktop extensions - -Voice mode - -Incognito chats - -User preferences - -Artifacts - -Projects - -Project sharing and collaboration - -Create and edit files with code execution - -Research - -Memory - -Skills - -Connectors - -Web search - -Enterprise search - -Claude in Chrome - -Claude for Microsoft 365 - -Claude for Microsoft Outlook - -Security and administration - -FeaturesTeamEnterprise (self-serve)Enterprise (sales-assisted) -Single sign-on (SSO) - -Domain verification - -Restrict org creation - -Migrate accounts using your domains - -Central billing and administration - -Admin controls for remote and local connectors - -Enterprise deployment for the Claude desktop app - -Usage analytics - -Organization wide skills deployment - -Domain capture - -Role-based access - -System for Cross-domain Identity Management (SCIM) - -Audit logs - -Compliance API - -Custom data retention controls - -Set org instructions - -HIPAA-ready offering - -Inference hooks - -Payment options - -FeaturesTeamEnterprise (self-serve)Enterprise (sales-assisted) -Credit card - -ACH - -Invoicing / net terms - -Billing cycle -Monthly and annual - -Annual - -Annual - -Adding seats midterm - -Partnership - -FeaturesTeamEnterprise (self-serve)Enterprise (sales-assisted) -AWS Marketplace availability - -Consultation - -Tiered incentives on committed spend - -Trials - -Customer success support at certain spend thresholds - -Models and usage - -FeaturesTeamEnterprise (self-serve)Enterprise (sales-assisted) -Fable - -Opus - -Sonnet - -Haiku - -Context window - -200k - -500k - -on default model -500k - -on default model - -Extra usage - -n/a - -n/a - -Priority access at high traffic times - -User and organizational level spend controls - -Model training - -None by default - -None by default - -None by default - -Latest models - -Contact salesContact sales - -Start buildingStart building - -Save 50% with batch processing. -Learn more -Batch processing - -Fable 5.1 -Next generation intelligence for long-running agents - -Input -$10 / MTok - -Output -$50 / MTok - -Prompt caching - -Read -$0.25 / MTok - -Write -$12.50 / MTok - -Opus 5 -Ideal for complex agentic coding and enterprise work - -Input -$5 / MTok - -Output -$25 / MTok - -Prompt caching - -Read -$0.50 / MTok - -Write -$6.25 / MTok - -Sonnet 5 -High-performance model for coding and agents - -Input -$2 / MTok - -Output -$10 / MTok - -Prompt caching - -Read -$0.20 / MTok - -Write -$2.50 / MTok - -Haiku 4.5 -Fastest, most cost-efficient model - -Input -$1 / MTok - -Output -$5 / MTok - -Prompt caching - -Read -$0.10 / MTok - -Write -$1.25 / MTok - -For workloads that need to run in the US, US-only inference is available at 1.1x pricing for input and output tokens. Learn more. - -Get up to 2.5x faster speeds with fast mode for Opus 5 at 2x standard pricing. Learn more. - -Prompt caching pricing reflects 5-minute TTL. Learn about extended prompt caching. - -Explore detailed pricingExplore detailed pricing - -Pricing for Claude Platform features - -Get more out of Claude with advanced features and capabilities. - -Learn moreLearn more - -Managed Agents - -Build and deploy agents at scale with a suite of composable APIs. Standard token rates apply. - -Cost - -$0.08 per session-hour for active runtime - -Web search - -Give Claude access to the latest information from the web. Doesn’t include input and output tokens required to process requests. - -Cost - -$10 / 1K searches - -Code execution - -Run Python code in a sandboxed environment for advanced data analysis. 50 free hours of usage daily per organization. - -Additional hours - -$0.05 per hour per container - -Service tiers - -Balance availability, performance, and predictable costs based on your needs. - -Learn moreLearn more - -Contact salesContact sales - -Standard - -Default tier for both piloting and scaling everyday use cases - -Batch - -For asynchronous workloads that can be processed together for better efficiency - -Legacy models - -Learn moreLearn more - -Explore detailed pricingExplore detailed pricing - -Fable 5 - -Input -$10 / MTok - -Output -$50 / MTok - -Prompt caching - -Read -$1 / MTok - -Write -$12.50 / MTok - -Opus 4.8 - -Input -$5 / MTok - -Output -$25 / MTok - -Prompt caching - -Read -$0.50 / MTok - -Write -$6.25 / MTok - -Save 50% with batch processing. -Learn more -Batch processing - -Sonnet 4.6 - -Input -$3 / MTok - -Output -$15 / MTok - -Prompt caching - -Read -$0.30 / MTok - -Write -$3.75 / MTok - -Opus 4.7 - -Input -$5 / MTok - -Output -$25 / MTok - -Prompt caching - -Read -$0.50 / MTok - -Write -$6.25 / MTok - -Opus 4.6 - -Input -$5 / MTok - -Output -$25 / MTok - -Prompt caching - -Read -$0.50 / MTok - -Write -$6.25 / MTok - -Sonnet 4.5 - -Input -$3 / MTok - -Output -$15 / MTok - -Prompt caching - -Read -$0.30 / MTok - -Write -$3.75 / MTok - -Opus 4.5 - -Input -$5 / MTok - -Output -$25 / MTok - -Prompt caching - -Read -$0.50 / MTok - -Write -$6.25 / MTok - -Opus 4.1 - -Input -$15 / MTok - -Output -$75 / MTok - -Prompt caching - -Read -$1.50 / MTok - -Write -$18.75 / MTok - -Prompt caching pricing reflects 5-minute TTL. Learn about extended prompt caching. - -FAQ - Plans and usage - -Billing and payments - -Managing your plan - -Active - -Plans and usage - What do I get with Claude Pro? -The Pro plan gives you everything in a Free plan with more usage and the full set of Claude capabilities. That includes Claude Code, Claude Cowork, Claude Design, along with projects to organize your chats and documents, access to more Claude models, and Claude for Microsoft 365. You can choose a monthly or annual subscription. - How is Claude Max different from Pro? -The Max plan is built for people who work with Claude throughout the day and want longer, uninterrupted sessions. You get everything in Pro, plus more usage: choose from 5x or 20x the usage of Pro per 5-hour session. Max also gives you higher output limits. Both options are billed monthly. See the cards above for current pricing. - What are the usage limits on each plan, and what happens when I hit them? -Every plan has usage limits that reset on a rolling five-hour session window, and paid plans add weekly limits on top. Your activity across Claude on web, desktop, mobile, and Claude Code all draws from the same pool. How much you can do depends on the length and complexity of your conversations, the model you choose, and the features you use, so there's no fixed message count. Free covers everyday questions. Pro gives you at least 5x more usage per 5-hour session than Free. Max gives you 5x or 20x more usage per 5-hour session than Pro. On Team plans, Standard seats give more than Pro and Premium seats give 5x more than Standard. To manage capacity and make sure all users have fair access, we may limit your usage in other ways, such as weekly and monthly caps or model and feature usage, at our discretion. When you reach a limit, you can wait for it to reset, move to a higher plan, or, on paid plans, turn on usage credits to keep working at standard API rates. You can see where you stand anytime in Settings > Usage. - Is Claude Code included in my plan? -Claude Code is included in all paid plans. It shares the same usage limits as the rest of your plan, so your work in the terminal and your chats draw from one pool. For heavy coding sessions, you can also switch to pay-as-you-go API credits through a Console account. - -How much does the Claude Enterprise plan cost? - -The Claude Enterprise plan is built for large businesses operating at scale. Pricing is $20 per seat per month plus usage billed at API rates, so cost scales with the models and tasks your team runs. Enterprise is billed annually and includes everything in Team plus advanced security and admin controls, including SCIM, audit logs, custom data retention, and role-based access. You can get started self-serve or contact sales for a tailored quote. - -PrevPrev - -NextNext - -Billing and payments - -Can I pay annually? Does it cost less than if I pay monthly? - -Yes, annual plans are designed to cost less than monthly subscriptions. Annual subscriptions are available for Pro and Team plans. Enterprise plans are only available as annual subscriptions. -When you choose to pay annually, you're billed once up front for the year. If you're already on a monthly plan, you can switch to annual in Settings > Billing, and any unused time is credited toward your first annual invoice. See the steps for Pro and Team. - -Does Claude offer refunds? - -Payments are generally non-refundable, except where our Consumer Terms of Service say otherwise or where local law requires it. -Customers outside of Europe can follow this process to request a refund: -Log in, open Get help from the menu, start a message, and choose Claude Refund Request to check your eligibility. -If you subscribed through the App Store, Apple handles the refund, so you'll request it from Apple. -If you subscribed through Google Play, contact our Support team to check eligibility for an active subscription. -If you're in the European Economic Area and the United Kingdom, and within your 14-day withdrawal period, you can request a refund directly in the app. - -PrevPrev - -NextNext - -Managing your plan - -How do I upgrade, downgrade, or switch between plans? - -You can change your plan any time in Settings. -For individual plans, go to Settings > Billing and choose Manage subscription. -On Team plans, an Owner or Primary Owner manages this under Organization settings > Billing. See more details here. -Upgrades take effect right away, and any unused time on your current plan is credited toward the new one, so you won't pay twice for the same days. -Moving from monthly to annual works the same way, with your remaining time credited to the first annual invoice. -In some upgrades (for example Pro Annual to Max 5x Monthly), your remaining value on your old plan is larger than the cost of one billing period of the new plan. In this case, the excess will be stored as credit, which will automatically get used for future purchases. -Downgrades take effect at the end of your current billing period, and your chats, projects, and files stay with your account whichever plan you're on. - -How do I cancel my Claude subscription? - -You can cancel anytime, and your plan stays active until the end of your current billing period. To avoid the next charge, cancel at least 24 hours before your renewal date. -If you signed up on the web or the desktop app, go to Settings > Billing and click Cancel. -If you subscribed through the App Store or Google Play, cancel from that app under Billing > Manage subscription, or directly in your Apple or Google Play account. -Team and self-serve Enterprise: an Owner or Primary Owner cancels under Organization settings > Billing. -Sales-assisted Enterprise: Reach out to your Anthropic Contact or our Sales team to discuss cancellation. -Canceling doesn't delete your data. Your chats, projects, and files stay with your account, though some features aren't available on the Free plan. For Pro and Max, your account moves to the Free plan once the period ends. For step-by-step help, see how to cancel a Pro or Max plan or cancel a Team plan. - -PrevPrev - -NextNext +How much does the Claude Enterprise plan cost?
| triage verdict | MATERIAL at 0.95, needed 0.55 |
|---|---|
| triage reasoning | Pricing page restructured with major plan tiers removed from public view and feature availability matrix reorganized. |
| signals | massive structural reorganization of pricing page · Team and Enterprise plan sections completely removed · Education plan removed · Legacy models section removed · FAQ content stripped · Individual plan structure replaced with simplified tier layout · Team plan pricing ($20/$100 per seat) no longer visible · Enterprise plan details removed from public pricing · model availability changed (Fable, Opus, Sonnet, Haiku availability per tier altered) · feature matrix restructured with Yes/No indicators added · context window specifications changed (200k shown for all tiers in new layout) · API pricing section completely removed from diff view |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Per-message effort changes are in beta on Claude Fable 5.1, Claude Mythos 5.1, and Claude Opus 5 on the Claude API. Add a role: "system" message with output_config.effort inside messages to change effort for later turns while preserving the prompt cache. Include the mid-conversation-output-config-2026-07-01 beta header in your requests. See Per-message effort. Turn-scoped system messages are in beta (mid-conversation-system-clear-at-2026-08-21 header). Set clear_at: "next_user_message" on a mid-conversation role: "system" message and it renders for the current turn only, then stays in the history at no token cost. Per-turn reminders don't accumulate and don't invalidate the prompt cache or later thinking blocks. thinking.display accepts a third value, "updates", in beta (thinking-display-updates-2026-08-18 header). Reasoning comes back with an empty thinking field, as under "omitted", and the short progress updates that Claude Fable 5.1, Claude Mythos 5.1, and Claude Fable 5 write between tool calls come back as text, at most one thinking block before a tool call. See Progress updates between tool calls. -Text generated by Claude Fable 5.1 and Claude Mythos 5.1 carries Anthropic's text watermark, and supported image and video files that Claude produces through the code execution tool carry C2PA Content Credentials when you retrieve them through the Files API on the Claude API. Marking requires no changes to your requests or response handling. +Text generated by Claude Fable 5.1 and Claude Mythos 5.1 carries Anthropic's text watermark, and supported image, video, and audio files that Claude produces through the code execution tool carry C2PA Content Credentials when you retrieve them through the Files API on the Claude API. Marking requires no changes to your requests or response handling. Like Claude Fable 5, both models require 30-day data retention and aren't available under zero data retention unless expressly authorized by Anthropic. See Model-specific data retention requirements. The guides for the Claude Enterprise endpoints of the Admin API (user management and spend limits), the Claude Enterprise Analytics API, and the Compliance API now show the anthropic-version header; send it on every request to these endpoints, as in the rest of the Claude API. See API versions.
| triage verdict | MATERIAL at 0.72, needed 0.45 |
|---|---|
| triage reasoning | Audio files now included in C2PA Content Credentials output—expands what Claude produces and marks through Files API. |
| signals | file type expansion in output credentials · data handling change (audio files now covered by C2PA) · new beta feature with header requirement · model capability change affecting output artifacts |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
claude-fable-5-1 1 Preview +2027-12-05 — -— claude-fable-5 1 @@ Suggest a fix? Last updated on -2026-08-27 +2026-09-02
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New future date (2027-12-05) added to claude-fable-5-1 entry signals model retirement or sunset event. |
| signals | future date added to model entry · date format suggests retirement/sunset deadline · model lifecycle change · last-updated timestamp changed |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Turn-scoped system messages are in beta (mid-conversation-system-clear-at-2026-08-21 header). Set clear_at: "next_user_message" on a mid-conversation role: "system" message and it renders for the current turn only, then stays in the history at no token cost. Per-turn reminders don't accumulate and don't invalidate the prompt cache or later thinking blocks. thinking.display accepts a third value, "updates", in beta (thinking-display-updates-2026-08-18 header). Reasoning comes back with an empty thinking field, as under "omitted", and the short progress updates that Claude Fable 5.1, Claude Mythos 5.1, and Claude Fable 5 write between tool calls come back as text, at most one thinking block before a tool call. See Progress updates between tool calls. Text generated by Claude Fable 5.1 and Claude Mythos 5.1 carries Anthropic's text watermark, and supported image and video files that Claude produces through the code execution tool carry C2PA Content Credentials when you retrieve them through the Files API on the Claude API. Marking requires no changes to your requests or response handling. -Like Claude Fable 5, both models require 30-day data retention and aren't available under zero data retention unless expressly authorized by Anthropic. See Model-specific data retention requirements. +Like Claude Fable 5, both models require 30-day data retention and aren't available under zero data retention unless expressly authorized by Anthropic. See Model-specific data retention requirements. +The guides for the Claude Enterprise endpoints of the Admin API (user management and spend limits), the Claude Enterprise Analytics API, and the Compliance API now show the anthropic-version header; send it on every request to these endpoints, as in the rest of the Claude API. See API versions.
| triage verdict | MATERIAL at 0.85, needed 0.45 |
|---|---|
| triage reasoning | New required header on Enterprise API endpoints is a breaking change for existing integrations. |
| signals | new API requirement: anthropic-version header now mandatory on three Enterprise endpoints · breaking change: Admin API, Analytics API, Compliance API affected · documentation update signals enforcement of existing versioning pattern |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Learn more Batch processing -Fable 5 +Fable 5.1 Next generation intelligence for long-running agents Input @@ Prompt caching +Read +$0.25 / MTok + Write $12.50 / MTok -Read -$1 / MTok - Opus 5 Ideal for complex agentic coding and enterprise work @@ Prompt caching -Write -$6.25 / MTok - Read $0.50 / MTok +Write +$6.25 / MTok + Sonnet 5 High-performance model for coding and agents @@ Prompt caching -Write -$2.50 / MTok - Read $0.20 / MTok +Write +$2.50 / MTok + Haiku 4.5 Fastest, most cost-efficient model @@ Prompt caching -Write -$1.25 / MTok - Read $0.10 / MTok +Write +$1.25 / MTok + For workloads that need to run in the US, US-only inference is available at 1.1x pricing for input and output tokens. Learn more. Get up to 2.5x faster speeds with fast mode for Opus 5 at 2x standard pricing. Learn more. @@ Explore detailed pricingExplore detailed pricing +Fable 5 + +Input +$10 / MTok + +Output +$50 / MTok + +Prompt caching + +Read +$1 / MTok + +Write +$12.50 / MTok + Opus 4.8 Input @@ Prompt caching -Write -$6.25 / MTok - Read $0.50 / MTok +Write +$6.25 / MTok + Save 50% with batch processing. Learn more Batch processing @@ Prompt caching -Write -$3.75 / MTok - Read $0.30 / MTok +Write +$3.75 / MTok + Opus 4.7 Input @@ Prompt caching -Write -$6.25 / MTok - Read $0.50 / MTok +Write +$6.25 / MTok + Opus 4.6 Input @@ Prompt caching -Write -$6.25 / MTok - Read $0.50 / MTok +Write +$6.25 / MTok + Sonnet 4.5 Input @@ Prompt caching -Write -$3.75 / MTok - Read $0.30 / MTok +Write +$3.75 / MTok + Opus 4.5 Input @@ Prompt caching -Write -$6.25 / MTok - Read $0.50 / MTok +Write +$6.25 / MTok + Opus 4.1 Input @@ Prompt caching -Write -$18.75 / MTok - Read $1.50 / MTok +Write +$18.75 / MTok + Prompt caching pricing reflects 5-minute TTL. Learn about extended prompt caching. FAQ
| triage verdict | MATERIAL at 0.95, needed 0.55 |
|---|---|
| triage reasoning | Fable 5.1 introduces new model with significantly lower cache read pricing. Fable 5 moved to legacy tier. |
| signals | model version bump: Fable 5 → Fable 5.1 · price change: Fable 5.1 read cache $1/MTok → $0.25/MTok (75% decrease) · new model tier added: Fable 5 appears in legacy section with different pricing · pricing reordering: read/write cache lines swapped (cosmetic) but prices are the substance · multiple prompt cache pricing changes across all models |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
2026-10-19 — +claude-opus-4-1 +— +Retired +2026-08-05 +claude-opus-5 + claude-haiku-4-5 1 GA @@ 2026-10-19 — +claude-fable-5-1 +1 +Preview +— +— + claude-fable-5 1 Preview 2027-12-05 — +claude-mythos-5-1 +1 +Preview +— +— + +claude-mythos-preview (gated research preview) +— +Preview +2027-04-02 +— + Cohere Model
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Retirement date for existing model and new model versions affect deployment planning and API compatibility. |
| signals | model retirement date added (claude-opus-4-1, 2026-08-05) · new model versions added to supported list (claude-opus-5, claude-fable-5-1, claude-mythos-5-1) · model status changes (Retired, Preview) · future sunset dates present |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — September 1, 2026 +https://platform.claude.com/docs/en/release-notes/overview#september-1-2026 +Tue, 01 Sep 2026 00:00:00 GMT + +We've launched Claude Fable 5.1 (claude-fable-5-1), the successor to Claude Fable 5 for long-running agentic coding, knowledge work, and research, alongside Claude Mythos 5.1 (claude-mythos-5-1) for Project Glasswing participants. Both models support a 1M token context window by default, 128k max output tokens, and always-on adaptive thinking, at $10 / $50 USD per MTok, the same as Claude Fable 5, with cache reads cut to $0.25 per MTok. Claude Fable 5.1 is available on the Claude API, Claude in Amazon Bedrock, Claude Platform on AWS, Claude on Google Cloud, and Claude in Microsoft Foundry. See What's new in Claude Fable 5.1 for capabilities, API changes, and migration guidance. +Prompt cache reads on Claude Fable 5.1 and Claude Mythos 5.1 cost $0.25 USD per million tokens: 0.025x the base input price, compared with 0.1x on other models. Cache writes are unchanged. See Prompt caching pricing. +On Claude Fable 5.1 and Claude Mythos 5.1, tool_choice types any and tool aren't supported and return a 400 error. auto and none are unchanged. To guarantee schema-conformant tool inputs, use strict tool use or structured outputs. +Thinking blocks produced by Claude Fable 5.1 and Claude Mythos 5.1 are preserved only for the model that produced them or a newer one: earlier models can't read them, and the API drops one replayed to an earlier model. Claude Fable 5.1 accepts thinking blocks from Claude Opus 5, Claude Fable 5, Claude Mythos 5, and earlier Claude models. On Claude Fable 5.1, the API also checks that nothing before a block has changed: for new accounts created on or after August 31, 2026, replaying one after the system prompt, tools, or an earlier message changed returns a 400 error. With the thinking-binding-controls-2026-08-01 beta header, dropped blocks are reported in an input_transformations response field, and thinking.block_binding.prefix_mismatch_behavior chooses between rejecting and dropping blocks whose history changed. See Preserved thinking. +Per-message effort changes are in beta on Claude Fable 5.1, Claude Mythos 5.1, and Claude Opus 5 on the Claude API. Add a role: "system" message with output_config.effort inside messages to change effort for later turns while preserving the prompt cache. Include the mid-conversation-output-config-2026-07-01 beta header in your requests. See Per-message effort. +Turn-scoped system messages are in beta (mid-conversation-system-clear-at-2026-08-21 header). Set clear_at: "next_user_message" on a mid-conversation role: "system" message and it renders for the current turn only, then stays in the history at no token cost. Per-turn reminders don't accumulate and don't invalidate the prompt cache or later thinking blocks. +thinking.display accepts a third value, "updates", in beta (thinking-display-updates-2026-08-18 header). Reasoning comes back with an empty thinking field, as under "omitted", and the short progress updates that Claude Fable 5.1, Claude Mythos 5.1, and Claude Fable 5 write between tool calls come back as text, at most one thinking block before a tool call. See Progress updates between tool calls. +Text generated by Claude Fable 5.1 and Claude Mythos 5.1 carries Anthropic's text watermark, and supported image and video files that Claude produces through the code execution tool carry C2PA Content Credentials when you retrieve them through the Files API on the Claude API. Marking requires no changes to your requests or response handling. +Like Claude Fable 5, both models require 30-day data retention and aren't available under zero data retention unless expressly authorized by Anthropic. See Model-specific data retention requirements.
| triage verdict | MATERIAL at 0.98, needed 0.45 |
|---|---|
| triage reasoning | New model release with pricing changes, breaking API changes, data retention requirements, and multiple feature additions affecting production deployments. |
| signals | new model release (Claude Fable 5.1, Claude Mythos 5.1) · pricing change (cache reads $0.25/MTok, 0.025x vs 0.1x) · breaking API change (tool_choice types any/tool return 400 error) · new data retention requirement (30-day minimum, no zero-retention option) · new feature (adaptive thinking always-on) · context window specification (1M tokens default, 128k max output) · beta API features with headers (mid-conversation output config, system message scoping, thinking display updates) · thinking block compatibility constraints (forward-compatible only, prefix mismatch validation for new accounts after Aug 31 2026) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Upcoming deprecations are listed below, with the most recent announcements at the top. +2026-08-26: Transcription models + +On August 26, 2026, we notified developers using whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe, and gpt-4o-transcribe-diarize of their deprecation and removal from the API on February 26, 2027. + +For information about the recommended replacements, see the transcription guide. + +Shutdown dateModel / systemRecommended replacement +Feb 26, 2027whisper-1gpt-live-transcribe or gpt-transcribe +Feb 26, 2027gpt-4o-transcribegpt-live-transcribe or gpt-transcribe +Feb 26, 2027gpt-4o-mini-transcribegpt-live-transcribe or gpt-transcribe +Feb 26, 2027gpt-4o-transcribe-diarizegpt-live-transcribe or gpt-transcribe + 2026-07-20: Legacy audio, realtime, and transcription models On July 20, 2026, we notified developers using legacy audio, realtime, and transcription model families and snapshots of their deprecation and removal from the API on January 20, 2027. @@ 2025-08-20: Assistants API -The Assistants API was officially sunset on August 26, 2026, following its deprecation announcement on August 26, 2025. +On August 26th, 2025, we notified developers using the Assistants API of its deprecation and removal from the API one year later, on August 26, 2026. When we released the Responses API in March 2025, we announced plans to bring all Assistants API features to the easier to use Responses API, with a sunset date in 2026.
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | New deprecation entry with four transcription models, shutdown date, and replacement guidance for engineering teams. |
| signals | future sunset date added (Feb 26, 2027) · four models deprecated and removed · replacement models specified · deprecation announcement date (Aug 26, 2026) · existing deprecation (Assistants API) reworded for clarity |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
2027-05-20 — +MAI-Image-2.5-Pro +2026-06-19 +Preview +2026-10-01 +— + +MAI-Image-2.5-Flash +2026-06-02 +Preview +2026-10-01 +— + +MAI-Image-2.5 +2026-06-02 +Preview +2026-10-01 +— + MAI-Image-2e 2026-04-09 -Preview +Retired 2026-08-15 MAI-Image-2.5-Flash
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | New models added with concrete deprecation and retirement dates; existing model status changed to Retired. |
| signals | new models added to retirement schedule · future deprecation dates (2026-06-02, 2026-06-19) · future sunset dates (2026-10-01) · model status change: Preview → Retired · three new model variants with defined lifecycle windows |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Anthropic +Microsoft Foundry offers Claude models in two versions: + +Version 1: Hosted on Anthropic infrastructure + +Version 2: Hosted on Azure + Model Version Lifecycle Retirement date Replacement -claude-haiku-4-5 +claude-sonnet-5 +2 +GA +2027-06-30 — -Preview -2026-10-19 + +claude-sonnet-5 +1 +GA +2027-06-30 — -claude-mythos-preview (gated research preview) +claude-sonnet-4-6 +1 +GA +2027-02-10 — -Preview -2027-04-02 + +claude-sonnet-4-5 +1 +GA +2026-10-19 — -claude-opus-4-1 +claude-opus-5 +2 +GA +2027-07-08 — -Retired -2026-08-05 + claude-opus-5 +1 +GA +2027-07-08 +— -claude-opus-4-5 +claude-opus-4-8 +2 +GA +2027-09-01 — -Preview -2026-10-19 -— -claude-opus-4-6 +claude-opus-4-8 +1 +GA +2027-09-01 — -Preview -2027-02-02 -— claude-opus-4-7 -— -Preview +1 +GA 2027-04-06 — -claude-sonnet-4-5 +claude-opus-4-6 +1 +GA +2027-02-02 — -Preview + +claude-opus-4-5 +1 +GA 2026-10-19 — -claude-sonnet-4-6 +claude-haiku-4-5 +1 +GA +2026-10-19 — + +claude-haiku-4-5 +2 +GA +2026-10-19 +— + +claude-fable-5 +1 Preview -2027-02-10 +2027-12-05 — Cohere @@ Suggest a fix? Last updated on -2026-08-26 +2026-08-27
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Retirement dates shifted, model availability split by hosting, lifecycle statuses upgraded, breaking changes to model roster. |
| signals | multiple retirement dates changed · model lifecycle statuses changed (Preview→GA) · new model versions introduced (Version 1 vs Version 2 hosting split) · models removed from list (claude-haiku-4-5 Preview variant gone) · new model added (claude-fable-5) · retirement date accelerated (claude-opus-4-1 from 2026-08-05 now absent) · structural change: dual-version hosting model introduced |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — August 26, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-26-2026 +Wed, 26 Aug 2026 00:00:00 GMT + +The Compliance API session endpoints are out of beta for Cowork and Claude Code sessions. See Retrieve session transcripts. +The Compliance API local session endpoints now also return transcripts of Claude Science sessions (product_surface value claude_science) and Claude for Microsoft 365 sessions in Excel, PowerPoint, Word, and Outlook (product_surface values beginning with office_agents), in beta for Claude Enterprise organizations, with your existing Compliance Access Key and the read:compliance_user_data scope. See Sessions on users' machines.
| triage verdict | MATERIAL at 0.78, needed 0.45 |
|---|---|
| triage reasoning | Compliance API endpoints graduate from beta and expand to new product surfaces, affecting API consumers' integration scope. |
| signals | feature graduation from beta to GA · new product surfaces added to API · new beta feature announced · API scope/capability expansion · compliance/data handling change |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
https://platform.claude.com/docs/en/release-notes/overview#august-20-2026 Thu, 20 Aug 2026 00:00:00 GMT -We've released v1.0 of the Python SDK. The SDK's HTTP layer moves from httpx to httpx2, a maintained, API-compatible fork: build custom http_client, Timeout, and transport objects from httpx2 (the DefaultHttpxClient helpers are unchanged), and call httpx2.alias_httpx() at startup if you rely on tracing or mocking libraries that patch httpx. v1.0 requires Python 3.10 or later and removes long-deprecated surface, including the legacy Text Completions API, the temperature, top_p, and top_k parameters on Messages methods, and the tool runner's client-side compaction_control. On the async client, .with_raw_response results now need await response.parse(), and AnthropicBedrock now raises an error when no AWS region is configured instead of defaulting to us-east-1. See the v1 migration guide for every change with before-and-after snippets. +We've released v1.0 of the Python SDK. The SDK's HTTP layer moves from httpx to httpx2, a maintained, API-compatible fork: build custom http_client, Timeout, and transport objects from httpx2 (the DefaultHttpxClient helpers are unchanged), and call httpx2.alias_httpx() at startup if you rely on tracing or mocking libraries that patch httpx. v1.0 requires Python 3.10 or later and removes long-deprecated surface, including the legacy Text Completions API, the temperature, top_p, and top_k parameters on Messages methods, and the tool runner's client-side compaction_control. On the async client, .with_raw_response results now need await response.parse(), and AnthropicBedrock now raises an error when no AWS region is configured instead of defaulting to us-east-1. See the v1 migration guide for every change with before-and-after snippets. +The computer use and browser use toolsets (computer_toolset_20260801 and browser_toolset_20260801) are now available on Google Cloud for Claude Fable 5, Claude Mythos 5, Claude Opus 5, Claude Sonnet 5, and Claude Opus 4.8. Requests use the same tools entries as on the Claude API.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New toolsets and models available on Google Cloud platform expands deployment options for engineering teams. |
| signals | new feature availability announced · specific model list provided · new toolsets with dated identifiers · platform expansion (Google Cloud) · affects deployment decisions |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
sora-2 2025-12-08 Preview -2026-09-15 +2026-10-15 — text-embedding-3-large
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Sora-2 preview retirement date pushed from September to October 2026. Teams relying on this model need to know the new deadline. |
| signals | future date changed · retirement/sunset date extended by one month · model lifecycle milestone altered |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
We've launched the MCP connector in public beta, a feature that allows you to connect to remote MCP servers directly from the Messages API. To increase answer quality and decrease tool errors, we've changed the default value for the top_p nucleus sampling parameter in the Messages API from 0.999 to 0.99 for all models. To revert this change, set top_p to 0.999. Additionally, when extended thinking is enabled, you can now set top_p to values between 0.95 and 1. -We've moved our Go SDK from beta to GA. +Our Go SDK has moved from beta to its first stable release. We've included minute and hour level granularity to the Usage page of Console alongside 429 error rates on the Usage page.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | Default API parameter change affects all users; SDK GA is a stability milestone; extended thinking constraint change expands capability. |
| signals | default parameter value changed (top_p 0.999 to 0.99) · parameter constraint relaxed (extended thinking top_p now 0.95–1.0) · SDK lifecycle milestone (beta to GA) · behavioral change affecting all models without opt-in |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — August 20, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-20-2026 +Thu, 20 Aug 2026 00:00:00 GMT + +We've released v1.0 of the Python SDK. The SDK's HTTP layer moves from httpx to httpx2, a maintained, API-compatible fork: build custom http_client, Timeout, and transport objects from httpx2 (the DefaultHttpxClient helpers are unchanged), and call httpx2.alias_httpx() at startup if you rely on tracing or mocking libraries that patch httpx. v1.0 requires Python 3.10 or later and removes long-deprecated surface, including the legacy Text Completions API, the temperature, top_p, and top_k parameters on Messages methods, and the tool runner's client-side compaction_control. On the async client, .with_raw_response results now need await response.parse(), and AnthropicBedrock now raises an error when no AWS region is configured instead of defaulting to us-east-1. See the v1 migration guide for every change with before-and-after snippets.
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Major SDK version with multiple breaking changes and removed APIs requires engineering team action. |
| signals | major version release (v1.0) · breaking API changes (legacy Text Completions API removed) · deprecated parameters removed (temperature, top_p, top_k) · minimum Python version requirement (3.10+) · async client behavior change (await response.parse()) · error behavior change (AnthropicBedrock region default removed) · dependency change (httpx to httpx2) · migration guide referenced |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
https://platform.claude.com/docs/en/release-notes/overview#august-19th-2024 Mon, 19 Aug 2024 00:00:00 GMT -We've moved 8,192 token outputs from beta to general availability for Claude Sonnet 3.5. +8,192-token outputs on Claude Sonnet 3.5 are out of beta and no longer require the max-tokens-3-5-sonnet-2024-07-15 header.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | Developers no longer need to pass a specific header to use 8K outputs on Sonnet 3.5. |
| signals | feature graduation from beta to GA · API header requirement removed · developer workflow change |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
https://platform.claude.com/docs/en/release-notes/overview#august-19-2026 Wed, 19 Aug 2026 00:00:00 GMT +The computer use tool is now generally available on the Claude API as the computer_toolset_20260801 toolset: no beta header, batch actions (several actions in one turn), zoom enabled by default, and per-member configuration through configs. Earlier beta versions remain available. Upgrading an existing integration changes the request shape and tool handling; see Migrate from computer_20251124. +We've launched the browser use tool (browser_toolset_20260801), a client toolset for driving a browser that your application hosts. It works inside a browser viewport rather than a whole desktop, reading the page itself (its accessibility tree, elements, forms, and tabs) and adding element references, form input, tab management, download reporting, and opt-in file upload on top of screenshot-and-click control. +Both toolsets are available for Claude Fable 5, Claude Mythos 5, Claude Opus 5, Claude Sonnet 5, and Claude Opus 4.8 on the Claude API. The Files API is now generally available on the Claude API. Requests to the /v1/files endpoints, and Messages API requests that reference an uploaded file, no longer require the files-api-2025-04-14 beta header. Requests sent without the header use the GA response format: file expiration (set expires_in_seconds when you upload a file; file objects report expires_at), and page and next_page pagination plus an ids[] filter when you list files. /v1/files requests that still send the beta header keep working and return the previous response format. Agent Skills and the Skills API (/v1/skills) are now generally available on the Claude API. Requests no longer require the skills-2025-10-02 beta header, including Messages API requests that load Skills through the container parameter. Requests that still send the header continue to work unchanged. See Using Agent Skills with the API. -The Admin API user-management endpoints for Claude Enterprise (claude.ai) organizations (members, invites, groups, and custom roles) are now generally available. The anthropic-beta: ce-user-management-2026-07-13 header is no longer required on group and custom-role requests; requests that still send it are accepted unchanged. See User management. +The Admin API user-management endpoints for Claude Enterprise (claude.ai) organizations (members, invites, groups, and custom roles) are now generally available. The anthropic-beta: ce-user-management-2026-07-13 header is no longer required on group and custom-role requests; requests that still send it are accepted unchanged. See User management. +You can now restrict which sites a Claude Managed Agents agent's web_search and web_fetch tools can reach. Set allowed_domains or blocked_domains on the tool's entry in the agent_toolset_20260401 configs array; web_fetch also accepts max_content_tokens and web_search accepts user_location. Each configs entry is identified by its name and typed by an optional type, and requests that pass only name, enabled, and permission_policy continue to work; in the typed SDKs, configs entries become per-tool types. See Restrict web search and web fetch domains. +Claude Managed Agents sessions that run in a self-hosted sandbox can now attach memory stores. The Python, TypeScript, and Go SDK workers download each attached store into the sandbox at its mount_path and sync the agent's changes back to the store. See Use memory stores. +The session viewer in the Claude Console has been redesigned with a timeline minimap, a transcript grouped by model request, and an Inspector panel for session details and cost, raw events, per-tool statistics, mounted resources, and per-thread activity. See Console observability.
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Multiple GA releases, new tools, breaking changes, and feature additions affecting API integrations. |
| signals | computer_toolset_20260801 GA release with breaking change warning · new browser_toolset_20260801 tool launched · Files API GA with response format changes · Agent Skills API GA · Admin API user-management GA · new web_search/web_fetch domain restriction feature · new memory store attachment for self-hosted sandbox sessions · multiple model availability statements (Fable 5, Mythos 5, Opus 5, Sonnet 5, Opus 4.8) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
+## Claude Platform release notes — August 19, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-19-2026 +Wed, 19 Aug 2026 00:00:00 GMT + +The Files API is now generally available on the Claude API. Requests to the /v1/files endpoints, and Messages API requests that reference an uploaded file, no longer require the files-api-2025-04-14 beta header. Requests sent without the header use the GA response format: file expiration (set expires_in_seconds when you upload a file; file objects report expires_at), and page and next_page pagination plus an ids[] filter when you list files. /v1/files requests that still send the beta header keep working and return the previous response format. +Agent Skills and the Skills API (/v1/skills) are now generally available on the Claude API. Requests no longer require the skills-2025-10-02 beta header, including Messages API requests that load Skills through the container parameter. Requests that still send the header continue to work unchanged. See Using Agent Skills with the API. +The Admin API user-management endpoints for Claude Enterprise (claude.ai) organizations (members, invites, groups, and custom roles) are now generally available. The anthropic-beta: ce-user-management-2026-07-13 header is no longer required on group and custom-role requests; requests that still send it are accepted unchanged. See User management.
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | Three major API features graduating from beta to GA with breaking response format changes and header removal. |
| signals | Files API moved from beta to GA · Skills API moved from beta to GA · Admin API user-management endpoints moved from beta to GA · Beta headers no longer required · Response format changes documented · Breaking change: GA response format differs from beta format · Future date: August 19, 2026 |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
gpt-4.1 2025-04-14 -Deprecated +Legacy 2027-04-14 — gpt-4.1-mini 2025-04-14 -Deprecated +Legacy 2027-04-14 — gpt-4.1-nano 2025-04-14 -Deprecated -2026-10-14 +Legacy +2027-04-14 — gpt-4o @@ gpt-4o 2024-11-20 -Deprecated +Legacy 2027-04-14 gpt-5.1
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | Model retirement dates extended and status reclassified; affects deployment planning and migration timelines. |
| signals | status label changed from 'Deprecated' to 'Legacy' · gpt-4.1-nano retirement date extended from 2026-10-14 to 2027-04-14 · multiple models affected · sunset dates are future dates paired with retirement timeline |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
2026-04-23 GA 2028-02-20 +DeepSeek-V4-Flash-0731 + +DeepSeek-V4-Flash-0731 +2026-07-31 +Preview +2026-12-03 — DeepSeek-V4-Pro @@ Suggest a fix? Last updated on -2026-08-17 +2026-08-19
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | New model added to public retirement schedule with defined preview and GA dates. |
| signals | new model added to retirement schedule · model name: DeepSeek-V4-Flash-0731 · preview end date: 2026-12-03 · GA date: 2026-07-31 · retirement schedule table expanded |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
DeepSeek-R1 1 -Legacy +Retired 2026-08-13 DeepSeek-V4-Pro @@ MAI-Image-2 2026-02-20 -Preview +Retired 2026-08-15 MAI-Image-2.5 +MAI-Transcribe-1 +2026-01-23 +Preview +2026-09-15 +MAI-Transcribe-1.5 + Mistral AI Model @@ Stable-Diffusion-3.5-Large 1 -Deprecated +Retired 2026-07-31 — Stable-Image-Core 1 -Deprecated +Retired 2026-07-31 — Stable-Image-Ultra 1 -Deprecated +Retired 2026-07-31 — @@ Suggest a fix? Last updated on -2026-07-24 +2026-08-17
| triage verdict | MATERIAL at 0.95, needed 0.45 |
|---|---|
| triage reasoning | Multiple models moved to Retired status with concrete sunset dates. New models added to retirement schedule. |
| signals | status label changes from Legacy/Preview/Deprecated to Retired · retirement dates present (2026-07-31, 2026-08-13, 2026-08-15, 2026-09-15) · new model entries added (MAI-Transcribe-1, MAI-Transcribe-1.5) · multiple models transitioned to final retirement state · documentation timestamp updated |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Tue, 11 Aug 2026 00:00:00 GMT The Compliance API now returns transcripts of Cowork and Claude Code sessions that run on your users' machines, in beta for Claude Enterprise organizations. GET /v1/compliance/apps/sessions/local lists sessions across your organization, GET /v1/compliance/apps/sessions/local/{session_id} retrieves one session's metadata, and GET /v1/compliance/apps/sessions/local/{session_id}/messages returns its transcript, all with your existing Compliance Access Key and the read:compliance_user_data scope. See Retrieve local sessions. +We've added the anthropic-workspace-id response header to the Claude API. It carries the wrkspc_-prefixed ID of the workspace that the request's API key or access token resolved to, including your organization's Default Workspace. See Identify the workspace behind an API response. ## Claude Platform release notes — August 12, 2025 https://platform.claude.com/docs/en/release-notes/overview#august-12-2025
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New API endpoints and response headers enable previously unavailable data access patterns for compliance use cases. |
| signals | new API endpoint added (GET /v1/compliance/apps/sessions/local) · new API endpoint added (GET /v1/compliance/apps/sessions/local/{session_id}) · new API endpoint added (GET /v1/compliance/apps/sessions/local/{session_id}/messages) · new response header added (anthropic-workspace-id) · new scope introduced (read:compliance_user_data) · beta feature for Enterprise tier · data access capability change (local session transcripts now retrievable) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
https://platform.claude.com/docs/en/release-notes/overview#april-14-2026 Tue, 14 Apr 2026 00:00:00 GMT -We announced the deprecation of the Claude Sonnet 4 model (claude-sonnet-4-20250514) and the Claude Opus 4 model (claude-opus-4-20250514), with retirement on the Claude API scheduled for June 15, 2026. We recommend migrating to Claude Sonnet 4.6 and Claude Opus 4.8 respectively. Read more in model deprecations. +We announced the deprecation of the Claude Sonnet 4 model (claude-sonnet-4-20250514) and the Claude Opus 4 model (claude-opus-4-20250514), with retirement on the Claude API scheduled for June 15, 2026. We recommend migrating to Claude Sonnet 4.6 and Claude Opus 4.8 respectively. Read more in Model deprecations. ## Claude Platform release notes — April 16, 2026 https://platform.claude.com/docs/en/release-notes/overview#april-16-2026 @@ Some customers might encounter 429 (rate_limit_error) errors following a sharp increase in API usage due to acceleration limits on the API. Previously, 529 (overloaded_error) errors would occur in similar scenarios. +## Claude Platform release notes — August 11, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-11-2026 +Tue, 11 Aug 2026 00:00:00 GMT + +The Compliance API now returns transcripts of Cowork and Claude Code sessions that run on your users' machines, in beta for Claude Enterprise organizations. GET /v1/compliance/apps/sessions/local lists sessions across your organization, GET /v1/compliance/apps/sessions/local/{session_id} retrieves one session's metadata, and GET /v1/compliance/apps/sessions/local/{session_id}/messages returns its transcript, all with your existing Compliance Access Key and the read:compliance_user_data scope. See Retrieve local sessions. + ## Claude Platform release notes — August 12, 2025 https://platform.claude.com/docs/en/release-notes/overview#august-12-2025 Tue, 12 Aug 2025 00:00:00 GMT @@ Wed, 05 Aug 2026 00:00:00 GMT Inference hooks are now in beta for Claude Enterprise organizations. Point Claude at your organization's AI security server, and each governed prompt across claude.ai, Cowork, and Claude Code is held for the server's allow or deny verdict before inference proceeds. Requests are signed, failure handling is configurable, and every denial is recorded in the compliance Activity Feed. See Inference hooks. -We've retired the Claude Opus 4.1 model (claude-opus-4-1-20250805). All requests to this model will now return an error. We recommend upgrading to Claude Opus 5. Researchers can request ongoing access through the External Researcher Access Program. +We've retired the Claude Opus 4.1 model (claude-opus-4-1-20250805). All requests to this model on the Claude API will now return an error. We recommend upgrading to Claude Opus 5. Researchers can request ongoing access through the External Researcher Access Program. +## Claude Platform release notes — August 7, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-7-2026 +Fri, 07 Aug 2026 00:00:00 GMT + +You can now set a budget on a Claude Managed Agents session: a hard cap on the session's spend, priced at public list rates. A session that reaches its budget pauses with the budget_reached stop reason instead of starting new model requests; changing or removing the budget resumes it. Deployments accept the same budget and apply it to each session they start. See Session budgets. +You can now give a Claude Managed Agents session an advisor: a model at least as capable as the agent's own that the session's primary thread can consult mid-turn for strategic guidance. Configure it as a {"type": "advisor"} entry in the agent's multiagent roster, naming the model to consult. See Give the session an advisor. +You can now control where model inference runs for a Claude Managed Agents agent. Set inference_geo inside the model object when you create the agent, or override it for a single session. See Data residency for the available geos and pricing. +Claude Managed Agents sessions can now load skills from a GitHub repository. When a session mounts a repository, any skills in its root .claude/skills directory are discovered automatically at session start and available to the agent for that session. + ## Claude Platform release notes — August 8, 2025 https://platform.claude.com/docs/en/release-notes/overview#august-8-2025 Fri, 08 Aug 2025 00:00:00 GMT @@ Thu, 19 Feb 2026 00:00:00 GMT We've launched automatic caching for the Messages API. Add a single cache_control field to your request body and the system automatically caches the last cacheable block, moving the cache point forward as conversations grow. No manual breakpoint management required. Works alongside existing block-level cache control for fine-grained optimization. Available on the Claude API and Microsoft Foundry (preview). Learn more in Prompt caching. -We've retired the Claude Sonnet 3.7 model (claude-3-7-sonnet-20250219) and the Claude Haiku 3.5 model (claude-3-5-haiku-20241022). All requests to these models will now return an error. We recommend upgrading to Claude Sonnet 4.6 and Claude Haiku 4.5 respectively. Researchers can request ongoing access through the External Researcher Access Program. +We've retired the Claude Sonnet 3.7 model (claude-3-7-sonnet-20250219) and the Claude Haiku 3.5 model (claude-3-5-haiku-20241022). All requests to Claude Sonnet 3.7 will now return an error. Requests to Claude Haiku 3.5 on the Claude API will now return an error; it remains available on Amazon Bedrock and Google Cloud. We recommend upgrading to Claude Sonnet 4.6 and Claude Haiku 4.5 respectively. Researchers can request ongoing access through the External Researcher Access Program. We announced the deprecation of the Claude Haiku 3 model (claude-3-haiku-20240307), with retirement scheduled for April 20, 2026. We recommend migrating to Claude Haiku 4.5. Read more in Model deprecations. ## Claude Platform release notes — February 24th, 2025 @@ https://platform.claude.com/docs/en/release-notes/overview#june-15-2026 Mon, 15 Jun 2026 00:00:00 GMT -We've retired the Claude Sonnet 4 model (claude-sonnet-4-20250514) and the Claude Opus 4 model (claude-opus-4-20250514). All requests to these models will now return an error. We recommend upgrading to Claude Sonnet 4.6 and Claude Opus 4.8 respectively. Researchers can request ongoing access through the External Researcher Access Program. +We've retired the Claude Sonnet 4 model (claude-sonnet-4-20250514) and the Claude Opus 4 model (claude-opus-4-20250514). All requests to these models on the Claude API will now return an error. We recommend upgrading to Claude Sonnet 4.6 and Claude Opus 4.8 respectively. Researchers can request ongoing access through the External Researcher Access Program. ## Claude Platform release notes — June 18, 2026 https://platform.claude.com/docs/en/release-notes/overview#june-18-2026 @@ https://platform.claude.com/docs/en/release-notes/overview#june-5-2026 Fri, 05 Jun 2026 00:00:00 GMT -We announced the deprecation of the Claude Opus 4.1 model (claude-opus-4-1-20250805), with retirement on the Claude API scheduled for August 5, 2026. We recommend migrating to Claude Opus 4.8. Read more in model deprecations. +We announced the deprecation of the Claude Opus 4.1 model (claude-opus-4-1-20250805), with retirement on the Claude API scheduled for August 5, 2026. We recommend migrating to Claude Opus 4.8. Read more in Model deprecations. ## Claude Platform release notes — June 9, 2026 https://platform.claude.com/docs/en/release-notes/overview#june-9-2026
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | Multiple new API capabilities, features, and endpoints added; model availability clarified across platforms. |
| signals | new API endpoints added (Compliance API for local sessions) · new feature: Claude Managed Agents session budgets with hard spend caps · new feature: Claude Managed Agents advisor model capability · new feature: data residency control (inference_geo) for agents · new feature: GitHub repository skill loading for agents · model retirement clarification (Claude Haiku 3.5 remains on Bedrock/GCP) · multiple new release notes entries (Aug 7, Aug 11, 2026) |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
Dreams (research preview) now supports Claude Opus 5. See Supported models. +## Claude Platform release notes — August 10, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-10-2026 +Mon, 10 Aug 2026 00:00:00 GMT + +The introductory pricing for Claude Sonnet 5 ($2 / $10 per MTok) is now the standard price: the previously scheduled increase to $3 / $15 per MTok on September 1, 2026 will not occur. See Pricing. + ## Claude Platform release notes — August 11, 2025 https://platform.claude.com/docs/en/release-notes/overview#august-11-2025 Mon, 11 Aug 2025 00:00:00 GMT @@ We launched a beta version of the PHP SDK. +## Claude Platform release notes — August 3, 2026 +https://platform.claude.com/docs/en/release-notes/overview#august-3-2026 +Mon, 03 Aug 2026 00:00:00 GMT + +The Compliance API now returns transcripts of Cowork sessions started on claude.ai web or mobile, in beta for Claude Enterprise organizations. GET /v1/compliance/apps/sessions/remote lists sessions and GET /v1/compliance/apps/sessions/remote/{session_id}/messages returns one session's transcript, using your existing Compliance Access Key with the read:compliance_user_data scope. See Retrieve remote sessions. + ## Claude Platform release notes — August 5, 2025 https://platform.claude.com/docs/en/release-notes/overview#august-5-2025 Tue, 05 Aug 2025 00:00:00 GMT @@ Wed, 05 Aug 2026 00:00:00 GMT Inference hooks are now in beta for Claude Enterprise organizations. Point Claude at your organization's AI security server, and each governed prompt across claude.ai, Cowork, and Claude Code is held for the server's allow or deny verdict before inference proceeds. Requests are signed, failure handling is configurable, and every denial is recorded in the compliance Activity Feed. See Inference hooks. +We've retired the Claude Opus 4.1 model (claude-opus-4-1-20250805). All requests to this model will now return an error. We recommend upgrading to Claude Opus 5. Researchers can request ongoing access through the External Researcher Access Program. ## Claude Platform release notes — August 8, 2025 https://platform.claude.com/docs/en/release-notes/overview#august-8-2025 @@ https://platform.claude.com/docs/en/release-notes/overview#june-30-2026 Tue, 30 Jun 2026 00:00:00 GMT -We've launched Claude Sonnet 5 (claude-sonnet-5), the next generation of our Sonnet model family, at introductory pricing of $2 / $10 per MTok through August 31, 2026 (standard $3 / $15 thereafter). Claude Sonnet 5 supports a 1M token context window, 128k max output tokens, and the same set of tools and platform features as Claude Sonnet 4.6, except Priority Tier, which is not available on Claude Sonnet 5. Three behavior changes apply when migrating: adaptive thinking is now on by default; manual extended thinking (thinking: {type: "enabled", budget_tokens: N}) is removed and returns a 400 error (it was deprecated on Sonnet 4.6); and setting sampling parameters (temperature, top_p, top_k) to non-default values returns a 400 error. Claude Sonnet 5 also uses a new tokenizer that produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. See What's new in Claude Sonnet 5 for details and migration guidance. For behavioral differences and model-specific prompting patterns, see Prompting Claude Sonnet 5. +We've launched Claude Sonnet 5 (claude-sonnet-5), the next generation of our Sonnet model family, at introductory pricing of $2 / $10 per MTok (made the standard price on August 10, 2026). Claude Sonnet 5 supports a 1M token context window, 128k max output tokens, and the same set of tools and platform features as Claude Sonnet 4.6, except Priority Tier, which is not available on Claude Sonnet 5. Three behavior changes apply when migrating: adaptive thinking is now on by default; manual extended thinking (thinking: {type: "enabled", budget_tokens: N}) is removed and returns a 400 error (it was deprecated on Sonnet 4.6); and setting sampling parameters (temperature, top_p, top_k) to non-default values returns a 400 error. Claude Sonnet 5 also uses a new tokenizer that produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. See What's new in Claude Sonnet 5 for details and migration guidance. For behavioral differences and model-specific prompting patterns, see Prompting Claude Sonnet 5. Claude Managed Agents session event streams now support event deltas. Opt in with the event_deltas[] query parameter on GET /v1/sessions/{session_id}/events/stream. The event_start and event_delta events preview an agent message's text as it's generated, before the complete agent.message event arrives. Listing sessions for Claude Managed Agents now supports backward pagination. GET /v1/sessions returns a prev_page cursor alongside next_page; pass it as the page parameter to return to the previous page. See Pagination. When creating a Claude Managed Agents session, you can now override the agent's configuration for that session. Pass agent with type: "agent_with_overrides" to replace the model, system prompt, tools, MCP servers, or skills for a single session. The agent itself is unchanged.
| triage verdict | MATERIAL at 0.98, needed 0.45 |
|---|---|
| triage reasoning | Pricing locked at lower tier, model sunset with migration path, new compliance API endpoints, and enterprise feature additions. |
| signals | pricing change: Claude Sonnet 5 introductory price ($2/$10 per MTok) made permanent; scheduled increase to $3/$15 cancelled · model retirement: Claude Opus 4.1 (claude-opus-4-1-20250805) retired, requests now error · future date with pricing impact: August 10, 2026 · new API endpoints: Compliance API now returns remote session transcripts (GET /v1/compliance/apps/sessions/remote) · new feature in beta: Inference hooks for Claude Enterprise |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |
- Batch requests receive a 50% discount on standard rates. + Batch requests receive a 40% discount on standard rates. + Effective 1 September 2026.
| triage verdict | MATERIAL at 0.95, needed 0.5 |
|---|---|
| triage reasoning | Batch pricing discount decreased materially with a future effective date. |
| signals | pricing change: discount reduced from 50% to 40% · future effective date: 1 September 2026 |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00154) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00646) |
- Requires Python 3.9 or later. + Requires Python 3.11 or later.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | Minimum Python version requirement raised from 3.9 to 3.11, forcing SDK users to upgrade their runtime environment. |
| signals | version floor increased · breaking change for existing deployments · affects SDK usability and compatibility |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00155) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00687) |
+ Customers on Free and Starter plans: inputs and outputs may be used + to improve our services. Enterprise plans are excluded.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | New clause explicitly permits using customer data to improve services on non-Enterprise plans, a material change to data handling terms. |
| signals | data handling terms changed · training data usage policy clarified · plan-level differentiation introduced · Enterprise exemption stated |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00158) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00660) |
- Inputs are retained for up to 30 days for abuse monitoring. + Inputs are retained as required for abuse monitoring.
| triage verdict | MATERIAL at 0.85, needed 0.5 |
|---|---|
| triage reasoning | Retention commitment shifted from bounded 30-day limit to open-ended 'as required,' materially altering data handling guarantees. |
| signals | data retention period softened from specific limit to indefinite · commitment weakened from 'up to 30 days' to 'as required' · change to data handling terms per channel rules |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00169) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00660) |
- Tier 2: 5,000 requests per minute + Tier 2: 10,000 requests per minute
| triage verdict | MATERIAL at 0.98, needed 0.45 |
|---|---|
| triage reasoning | Rate limit increase directly affects engineering capacity planning and API quota allocation. |
| signals | rate limit changed · numeric value doubled · tier capacity increased |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00148) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00610) |
- | model-v2-mini | generally available |
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | A model has been removed from the available offerings, affecting deployment options. |
| signals | model removed from supported list · removal from generally available tier |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00143) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00633) |
+ | legacy-completions | 2026-11-04 | See migration guide |
| triage verdict | MATERIAL at 0.92, needed 0.45 |
|---|---|
| triage reasoning | New deprecation with concrete sunset date signals actionable timeline for engineering teams. |
| signals | future sunset date · new deprecation entry · specific end-of-life date (2026-11-04) · migration guidance referenced |
| suppression | survived all deterministic rules |
| triage model | anthropic/claude-haiku-4.5 ($0.00156) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00651) |