DeepSeek V4 Flash completion tokens increased 25%
OpenRouter raised the completion token price for deepseek/deepseek-v4-flash-latest from $0.0000008 to $0.000001 per token, a 25% increase.
Method: on 2026-10-30 we checked a random sample of 50 items from the last 30 days against their sources. 43 were right, 2 were wrong, and 5 could not be checked because the source no longer shows the old data. The 95.6% figure leaves those 5 out; the strict 86% figure counts them as wrong. This sample was checked by RADAR's build assistant and was not independently double-checked by a person. With a sample this size the true rate is likely between 85.2% and 98.8%. One sample, not a guarantee.
OpenRouter raised the completion token price for deepseek/deepseek-v4-flash-latest from $0.0000008 to $0.000001 per token, a 25% increase.
OpenRouter's deepseek-pro-latest model pricing changed: prompt tokens from $0.000000558624 to $0.000000624 per token (+11.7%), completion tokens from $0.000001675872 to…
The top_provider.max_completion_tokens value increased from 32768 to 235929 tokens.
The thinking-binding-controls-2026-08-01 beta header adds a thinking_mismatch_allowed entry type to input_transformations responses.
The claude.com/pricing page was alternating between two layout variants, and RADAR read one swing as a change.
The claude.com/pricing page was alternating between two layout variants, and RADAR read one swing as a change.
On the Azure Foundry model retirement schedule, gpt-realtime-2.1 and gpt-realtime-2.1-mini, whose Version column reads 2026-07-07, changed lifecycle status from Preview…
The claude.com/pricing page was alternating between two layout variants, and RADAR read one swing as a change.
The retirement date for gpt-4.1-nano moved from 2026-10-14 to 2027-04-14, a six-month extension.
Max 5x and Max 20x tiers now offer monthly billing only, down from 'Monthly and annual'.
gpt-4.1-nano retirement moved from 2027-04-14 to 2026-10-14.
The Messages API now supports a `compaction` parameter (beta header `compact-2026-09-04`) that returns a signed compaction block summarizing sent messages.
OpenAI announced on September 11, 2026 that gpt-5.4-cyber is deprecated and will be removed from the API on October 1, 2026.
o1-pro, o3, o3-deep-research, and o3-pro moved from GA to Deprecated status.
Claude Managed Agents now supports an 'auto' permission policy mode where the server evaluates each tool call and either runs it, denies it, or pauses for approval.
gpt-image-2.5-flare and gpt-image-2.5-sunburst reached GA on 9 September 2026 and are scheduled for retirement on 9 September 2027.
Per-message effort changes (beta) are now supported on Google Cloud for Claude Fable 5.1, Claude Mythos 5.1, and Claude Opus 5.
The pricing comparison table no longer lists 'Access to Research' as a Pro tier feature.
The pricing page no longer displays Team plan pricing ($20-$25/seat standard, $100-$125/seat premium) or Enterprise plan details ($20/seat + API usage).
As of September 1, 2026, audio files generated by Claude Fable 5.1 and Claude Mythos 5.1 through the code execution tool now carry C2PA Content Credentials when retrieve…
Azure Foundry added a retirement date of 2027-12-05 for the claude-fable-5-1 model.
The Admin API (user management and spend limits), Analytics API, and Compliance API now require the anthropic-version header on every request, matching the existing requ…
Anthropic released Fable 5.1, replacing Fable 5 in the current model lineup.
claude-opus-4-1 marked Retired with effective date 2026-08-05.
Claude Fable 5.1 and Mythos 5.1 launched September 1, 2026.
On August 26, 2026, OpenAI announced that whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe, and gpt-4o-transcribe-diarize will be removed from the API on February 26…
MAI-Image-2.5, MAI-Image-2.5-Flash, and MAI-Image-2.5-Pro added to Azure Foundry retirement schedule.
Azure Foundry now offers Claude models in two versions: Version 1 (Anthropic infrastructure) and Version 2 (Azure infrastructure).
Compliance API session endpoints for Cowork and Claude Code are now GA (out of beta).
The computer_toolset_20260801 and browser_toolset_20260801 toolsets are now available on Google Cloud for Claude Fable 5, Claude Mythos 5, Claude Opus 5, Claude Sonnet 5…
The retirement date for the sora-2 preview model moved from 15 September 2026 to 15 October 2026.
The Messages API now defaults top_p to 0.99 instead of 0.999 across all models, effective immediately.
Released August 20, 2026.
The max-tokens-3-5-sonnet-2024-07-15 header is no longer required to access 8,192-token outputs on Claude Sonnet 3.5.
The computer use tool is now GA as computer_toolset_20260801, replacing earlier beta versions.
Effective August 19, 2026, three Claude APIs moved from beta to general availability: Files API (/v1/files), Skills API (/v1/skills), and Admin API user-management endpo…
Azure Foundry moved gpt-4.1-nano retirement from 2026-10-14 to 2027-04-14.
Azure added DeepSeek-V4-Flash-0731 to the retirement schedule.
DeepSeek-R1, MAI-Image-2, Stable-Diffusion-3.5-Large, Stable-Image-Core, and Stable-Image-Ultra moved from Legacy/Preview/Deprecated to Retired status between 31 July an…
The Claude API now returns an anthropic-workspace-id header in all responses, containing the wrkspc_-prefixed workspace ID that the request's API key or access token res…
Three new Compliance API endpoints launched August 11, 2026 in beta for Enterprise organizations: GET /v1/compliance/apps/sessions/local lists all local sessions, GET /v…
Claude Sonnet 5 pricing remains at $2 input / $10 output per MTok indefinitely.