// AI Platform Drift
Claude Fable 5.1 breaks tool_choice any/tool, requires 30-day retention
Claude Platform release notes · published 2026-09-01 20:40 UTC
WHAT CHANGEDClaude Fable 5.1 and Mythos 5.1 launched September 1, 2026. tool_choice types 'any' and 'tool' now return 400 errors; only 'auto' and 'none' work. Cache read pricing dropped from 0.1x to 0.025x base price ($0.25/MTok). Both models enforce 30-day minimum data retention with no zero-retention option unless Anthropic authorizes it. New accounts created August 31, 2026 or later get 400 errors when replaying thinking blocks after system prompt, tools, or message history changes.
WHY IT MATTERSIf you use tool_choice='any' or tool_choice='tool' in production, requests to Fable 5.1 will fail immediately. If your contract or compliance posture requires zero data retention, you cannot use these models without explicit authorization. The thinking block validation will break conversation replay for new accounts if you modify context mid-conversation.
WHAT TO DOGrep your codebase for tool_choice parameters set to 'any' or 'tool' and migrate to structured outputs or strict tool use before switching to Fable 5.1. Verify your data retention agreement permits 30-day retention or request authorization from Anthropic if you need zero retention. If you create new accounts after August 31, audit any code that replays thinking blocks after modifying system prompts or tools.
ACT-BY DATE PASSED — this item's act-by date, 2026-09-01, is before this page was generated 2026-11-21 (UTC). Check the primary source before acting on it.
Model-reported confidence 95% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## 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.
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| 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) |