Claude Fable 5.2, released by Anthropic on November 9, 2026, is its most capable generally available model. Sold as claude-fable-5-2, it costs $8 per million input tokens and $40 per million output tokens, down from $10 and $50 for Fable 5.1, keeps a 1M-token context window, and generates output up to 30% faster.
Last updated: November 9, 2026
Table of contents
- What is new in Claude Fable 5.2
- Specs, pricing and availability
- Fable 5.2 vs. Fable 5.1 and Opus 5.5
- How to migrate from Fable 5.1
- Effort settings and prompting patterns
- Limitations
- Frequently asked questions
- Sources
What is new in Claude Fable 5.2
Fable sits above Opus in Anthropic’s lineup. The Fable 5.2 and Mythos 5.2 announcement describes it as built for “the most daunting technical challenges” and says Fable and Mythos are the same underlying model. Fable carries cybersecurity and biology safeguards that let Anthropic offer it to all users, while Mythos 5.2 goes only to verified organizations.
The changes that matter most to power users:
- Lower price. $8 input and $40 output per million tokens, compared with $10 and $50 for Fable 5.1. CNBC describes this as a 20% cut.
- Faster output. Anthropic says outputs arrive up to 30% faster than Fable 5.1, and an optional fast mode speeds generation by up to 2.5 times.
- Better token efficiency. At medium effort, Anthropic says Fable 5.2 finishes hard tasks with fewer tokens than Fable 5.1. On FrontierCode 2.0, Fable 5.2 at high effort scored about the same as Opus 5.5 at max effort (45.3% vs. 45.5%) for roughly half the cost per task.
- New API rules. Preserved thinking now applies to every account, and the older computer use tool is gone. Details are in the migration section below.
Specs, pricing and availability
All figures below come from the Claude Fable 5.2 model page in the Claude Platform documentation.
| Item | Claude Fable 5.2 |
|---|---|
| API model ID | claude-fable-5-2 |
| Context window / max output | 1M tokens / 128K tokens |
| Input and output | Text and images in, text out |
| Thinking | Adaptive, always on; default effort high |
| Base price | $8 input, $40 output per 1M tokens |
| Prompt caching | $10 (5-minute write), $16 (1-hour write), $0.20 (read) per 1M tokens |
| Batch API | 50% off input and output |
| Fast mode | $16 input, $80 output per 1M tokens |
| Knowledge cutoff | June 2026 |
| Platforms | Claude API, Claude Platform on AWS, Claude Models on Google Cloud, Microsoft Foundry |
In the Claude apps, the Claude Fable product page lists availability on the Pro, Max, Ultra, Team and Enterprise plans. For launch week, Anthropic says eligible paid plans get one to three Fable usage resets, and Fable can use up to 50% of weekly limits on Pro plans and standard Team seats. US-only inference costs 1.1 times the standard rates. For how Fable compares with the cheaper tiers, see our Claude API pricing explainer and Claude Max cost breakdown.
Fable 5.2 vs. Fable 5.1 and Opus 5.5
These are Anthropic’s own scores at each model’s highest available effort, with safeguards on, so treat them as vendor-reported.
| Benchmark | Fable 5.2 | Fable 5.1 | Opus 5.5 |
|---|---|---|---|
| FrontierCode 2.0 (agentic coding) | 49.1% | 32.4% | 45.5% |
| FrontierSWE v2 (agentic coding) | 69.9% | 56.3% | 62.3% |
| AutomationBench (business workflows) | 51.4% | 31.4% | 42.5% |
| OSWorld 2.1, partial (computer use) | 89.3% | 83.5% | 87.2% |
| Humanity’s Last Exam (with tools) | 74.2% | 65.6% | 67.7% |
| Remote Labor Index | 41.3% | 17.9% | 21.3% |
Anthropic still recommends Opus 5.5 as the starting point for most coding and knowledge work, at $4 and $20 per million tokens. Both models share a 1M-token context, as our Claude context window guide explains. On the cheap end, Claude Haiku 5.5 remains the option for routing and extraction.
How to migrate from Fable 5.1
The What’s new in Claude Fable 5.2 page lists four breaking changes. Work through them in this order:
- Change the model ID from
claude-fable-5-1toclaude-fable-5-2in your requests or SDK constants. - If your code builds the
messagesarray itself, test for history edits. Changing the system prompt, tools or an earlier message before a thinking block now returns a 400 error on every account. As a stopgap, thethinking-binding-controls-2026-08-01beta header withprefix_mismatch_behavior: "drop_block"drops the affected blocks instead. - Replace any
computer_20251124tool with thecomputer_toolset_20260801toolset. - If you use the advisor tool with a Fable 5.2 executor, set the advisor model to
claude-fable-5-2orclaude-mythos-5-2, in new conversations. - Remove
budget_tokens,thinking: disabled, forcedtool_choicevalues and sampling parameters such astemperature, which all return 400 errors. - Re-run your evals with an effort sweep that starts at the default,
high.
Anthropic notes that nothing changes if Claude Code, claude.ai, Claude Managed Agents or the Claude Agent SDK builds your requests. Code that keeps system and tools fixed and only appends messages is also safe. Our guide to Claude message threads covers a related way to keep history append-only.
Effort settings and prompting patterns
Fable 5.2 supports five effort levels: low, medium, high, xhigh and max. Anthropic’s guidance is to start at high, step up to xhigh only for long coding and agent runs where you have measured a gain, and keep max for testing the ceiling.
- Lower effort for easy turns. Prompt lines such as “answer directly” no longer stop thinking at medium and above. Use a lower effort level instead.
- Ask it to finish. At low and medium effort, a turn can end with a list of remaining work. State clearly that the whole task should be completed.
- Say who owns the code. Without that, Anthropic found the model tends to report nearby problems instead of fixing them.
- Use it as an advisor or orchestrator. Anthropic suggests letting Haiku, Sonnet or Opus consult Fable 5.2 on hard problems, or having Fable 5.2 delegate simpler tasks to smaller models. Watch subagent costs, which can be a large share of a session.
Limitations
- Data retention. Fable 5.2 requires 30-day retention for safety monitoring and is not available under zero data retention unless Anthropic expressly authorizes it.
- No forced tool use. Use
tool_choice: autowith strict tool use or JSON outputs for schema-valid results. - Refusals and fallback. Declined requests return
stop_reason: "refusal". Permitted fallback targets are Claude Opus 4.8 and Claude Opus 5, and they cannot read Fable 5.2 thinking blocks.
Frequently asked questions
When was Claude Fable 5.2 released?
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026. Fable 5.2 is generally available that day in the Claude apps and on the Claude API as claude-fable-5-2. Anthropic says the model will not be retired sooner than November 9, 2027.
How much does Claude Fable 5.2 cost?
On the API, Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens. Cache reads cost $0.20, 5-minute cache writes $10 and 1-hour cache writes $16 per million tokens. Batch processing halves input and output prices, and fast mode costs $16 and $80.
Can I use Claude Fable 5.2 in Claude Code?
Yes. Anthropic says Fable 5.2 is available to Pro, Max, Ultra, Team and Enterprise users, and that fast mode for Fable 5.2 is available in Claude Code. Medium effort is the default in the Claude apps and Claude Code.
What is the difference between Claude Fable 5.2 and Claude Mythos 5.2?
Anthropic says Fable and Mythos are the same underlying model with the same specs and prices. Fable adds strong cybersecurity and biology safeguards so it can be offered to everyone. Mythos 5.2 is limited to organizations verified through Anthropic’s Cyber or Life Sciences Verification Programs.
Should I use Claude Fable 5.2 or Claude Opus 5.5?
Anthropic recommends starting with Claude Opus 5.5 for most coding and knowledge work, at half the price ($4 and $20 per million tokens). Move to Fable 5.2 for demanding reasoning and long-running agent work, or when your evals on Opus 5.5 at higher effort still fall short.
Sources
- Anthropic: Introducing Claude Fable 5.2 and Claude Mythos 5.2 (November 9, 2026)
- Claude Platform Docs: Claude Fable 5.2 model overview
- Claude Platform Docs: What’s new in Claude Fable 5.2
- Anthropic: Claude Fable product page
Image: announcement graphic from Anthropic’s Fable 5.2 and Mythos 5.2 announcement, used for editorial purposes. FreeAnswers is not affiliated with Anthropic.