Claude Fable 5.1: The Model Anthropic Tell You to Escalate To

What Claude Fable 5.1 is

Claude Fable 5.1 was released on 1 September 2026, and Anthropic's own documentation describes its role more precisely than any marketing line could. Their guidance is to start with Claude Opus 5.5 for most workloads, and to reach for Fable 5.1 for demanding reasoning and long-horizon agentic work — or, in their words, when your evaluations on Opus 5.5 at higher effort still fall short. It carries a 1 million token context window, returns up to 128,000 output tokens, takes text and image input, and runs adaptive thinking that is always on with a default effort of high. Thinking cannot be switched off.

Pricing and caching

At $10 per million input tokens and $50 per million output, Fable 5.1 is two and a half times the price of Opus 5.5. Anthropic's standard cache hit costs 10% of the base input price; on Fable 5.1 and Mythos 5.1 it costs 2.5%, or $0.25 per million tokens — the cheapest cache ratio in the lineup, and four times better in relative terms than the rest of it. The Batch API takes a further 50% off, to $5 per million input tokens and $25 output. Claude Mythos 5.1 is listed at the same rates and the same 2.5% cache ratio. There is no ongoing free tier, only a small trial credit for new accounts.

Independent figures

On LMArena's text leaderboard Claude Fable 5.1 has 1504 Elo (rank 5, max effort); by category, 1537 Elo in Coding (rank 15, max effort), 1528 Elo in Math (rank 2, max effort), 1533 Elo in Expert (rank 10, max effort) and 1489 Elo in Creative Writing (rank 8, max effort). SimpleQA Verified records 71.

Fable 5.1 and Fable 5

Anthropic's documentation directs most workloads to Opus 5.5 and names Fable 5.1 for demanding reasoning and long-horizon agentic work. Against Claude Fable 5, list prices are identical; Fable 5.1 charges 2.5% of base input for a cache hit where Fable 5 charges 10%. On LMArena, Fable 5 has 1504 Elo (rank 6, high effort).

Model IDs, platforms and lifecycle

The model ID is claude-fable-5-1 on the Claude API, Google Cloud, Microsoft Foundry and Claude Platform on AWS, and anthropic.claude-fable-5-1 on Amazon Bedrock. Anthropic's models overview gives a reliable knowledge cutoff of June 2026 and commits to retirement not sooner than 1 September 2027. Like every current Claude model it takes text and image input, returns text, and supports multilingual use, vision and tool use; the Models API returns its max_input_tokens, max_tokens and a capabilities object programmatically.

Cache writes, residency and the tokenizer

A 5-minute cache write costs $12.50 per million tokens and a 1-hour cache write $20, against the $0.25 cache read. Specifying US-only inference through the inference_geo parameter applies a 1.1x multiplier to every token category, including cache reads and writes; global routing, the default, uses standard pricing. On Amazon Bedrock and Google Cloud, regional and multi-region endpoints carry a 10% premium over global endpoints. Anthropic note that Claude 4.7 and later models use a newer tokenizer that produces approximately 30% more tokens for the same text, which applies to Fable 5.1's per-token prices.

Official Resources