Claude Fable 5.2 (Fast)

Claude Fable 5.2, released on November 9, 2026, is Anthropic's frontier model for demanding reasoning and long-running agent work, combining a large context window with improved coding, research, and visual reasoning capabilities. Its standard API pricing is $8 per million input tokens and $40 per million output tokens, so developers weighing an upgrade must judge whether the capability gains justify the cost on completed work. The page breaks down features including long-horizon agentic reasoning, stronger coding and software engineering, faster token generation, and scientific and mathematical reasoning, then covers benchmark results such as FrontierCode 2.0 and FrontierSWE v2 alongside Artificial Analysis comparisons. It closes with a migration checklist covering thinking-block prefix validation, model-bound thinking blocks, updating the computer-use tool, and reviewing data retention and request parameters.

Anthropic released Claude Fable 5.2 on November 9, 2026, making it generally available through the Claude API, Claude apps, Claude Code, Google Cloud, Microsoft Foundry, and Claude Platform on AWS. The model costs $8 per million input tokens and $40 per million output tokens, reduced from Fable 5.1’s $10 and $50 rates. Fable 5.2 provides a 1 million-token context window and a 128,000-token maximum output. Its fast mode costs $16 per million input tokens and $80 per million output tokens while generating tokens up to 2.5 times faster. Anthropic reports a 49.1% FrontierCode 2.0 score, compared with 32.4% for Fable 5.1 and 45.5% for Opus 5.5.

Claude Fable 5.2, released November 9, 2026, is listed at $8 per million input tokens and $40 per million output tokens, 20% less than Fable 5.1 and exactly twice the price of Opus 5.5. Cache reads cost $0.20 per million, batch pricing halves input and output rates, and fast mode doubles them to $16 and $80. The model ID is claude-fable-5-2 on the Claude API, Google Cloud, Microsoft Foundry, and Claude Platform on AWS, with anthropic.claude-fable-5-2 on Amazon Bedrock. The page also details a 1M token context window at a flat rate for every request size, 128K maximum output, always-on adaptive thinking, and a June 2026 knowledge cutoff, plus a 1.1x multiplier on all meters for US-only inference. It notes that Anthropic measured Fable 5.2 at High effort matching Opus 5.5 at Max for roughly half the cost per task, while warning that default effort, lower rate limits, and plan rules are the practical traps for teams.

Claude Fable 5.2 launched on November 9, 2026 as Anthropic’s strongest generally available model, alongside the restricted Claude Mythos 5.2 variant. Developers can select the public model through the Claude apps, Claude Code, and the API using the model ID claude-fable-5-2, with availability also reported on AWS, Google Cloud, and Microsoft Azure. The model supports a 1 million-token context window and up to 128,000 output tokens per request. Anthropic says it writes up to 30% faster than Fable 5.1, while its reported agentic coding score is 49.1%. Standard Fable 5.2 pricing is $8 per million input tokens and $40 per million output tokens, 20% below Fable 5.1.

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026, describing them as two names for the same underlying model, with Fable 5.2 carrying cyber and biology safeguards and Mythos 5.2 available only to vetted organizations through Anthropic's Cyber Verification or Life Sciences Verification Program. Fable 5.2 is broadly available on all platforms, and Mythos 5.2 is included in Claude Security for Claude Enterprise customers. Fable 5.2 generates outputs up to 30% faster than Fable 5.1 and uses fewer tokens than Fable 5.1 at Medium effort, the default in Claude apps and Claude Code, for challenging tasks. It can act as an Advisor that other Claude models consult or as an Orchestrator delegating simpler work to smaller models, and Anthropic's preserved-thinking documentation must be consulted to update integrations for the new API features.

Claude Fable 5.2 fast mode is described as a premium inference option on the same underlying model rather than a compressed mini variant, generating output up to 2.5 times faster than standard mode at $16 per million input tokens and $80 per million output tokens, exactly twice standard Fable 5.2 rates. The speedup targets output tokens per second once generation starts, with no claimed improvement to time to first token. The page notes that fast mode does not unlock a smarter reasoning system, so teams need no separate quality policy, though regression tests are still advised. It is not available with the Batch API or Priority Tier, requiring a separate route for those designs. Suggested uses include live coding and repair loops, customer-facing research agents, and time-sensitive operational assistance such as incident response.