Anthropic · 2026-09-01 · seismic
Claude Fable 5.1 — Anthropic's new top model, with cache reads 75% cheaper
Claude Fable 5.1 is Anthropic's new top-end model for coding, knowledge work and long-running agent tasks. It scores 55.8% on Terminal-Bench 4.0 against 42.0% for Claude Fable 5, and cache reads drop to $0.25 per million tokens.
Anthropic's new top-end Claude keeps the same $10/$50 rates but cuts cache reads by 75%, and jumps on long agent tasks.
Quick facts
| Maker | Anthropic |
|---|---|
| Released | 1 September 2026 |
| API model ID | claude-fable-5-1 |
| Context window | 1M tokens (128K max output) |
| Thinking | Adaptive, always on; default effort high |
| Mythos 5.1 access | Invitation only, through Project Glasswing |
| Availability | Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry |
Benchmarks
| Claude Fable 5.1 | 55.8% | |
|---|---|---|
| Claude Opus 5 | 52.3% | |
| Claude Fable 5 | 42% | |
| GPT-5.6 Sol | 37.3% |
| Claude Fable 5.1 | 52.6% | |
|---|---|---|
| Claude Opus 5 | 29% | |
| Claude Fable 5 | 24.7% | |
| GPT-5.6 Sol | 22.4% |
| Claude Fable 5.1 | 31.4% | |
|---|---|---|
| Claude Opus 5 | 26.9% | |
| GPT-5.6 Sol | 19.6% | |
| Claude Fable 5 | 17.1% |
Pricing
| Input | $10.00 / 1M tokens |
|---|---|
| Output | $50.00 / 1M tokens |
| Cache read · 75% less than Claude Fable 5 | $0.25 / 1M tokens |
| 5-minute cache write | $12.50 / 1M tokens |
| 1-hour cache write | $20.00 / 1M tokens |
| Batch API · Applies to input and output | 50% off |
What is it?
Claude Fable 5.1 replaces Claude Fable 5 as Anthropic's most capable model, built for demanding reasoning and long-horizon agentic work. It ships with the same $10 input and $50 output rates per million tokens, a 1M-token context window and 128K max output. Claude Mythos 5.1 is the same underlying model with lighter safeguards, and stays invitation-only.
How does it work?
Thinking is adaptive and always on, steered by an effort parameter that defaults to high, and a beta lets you change effort part-way through a conversation without losing the prompt cache. The gains show up most on long jobs: Terminal-Bench-Science 0.1 goes from 24.7% on Fable 5 to 52.6%, and AutomationBench from 17.1% to 31.4%. Three API behaviours break for callers moving off Fable 5.
Why does it matter?
Long agent runs re-read the same cached context on every step, so the 75% cut on cache reads lands where the bill actually is. Anthropic puts the saving at about 25% for typical workloads and up to about 45% for complex coding and highly agentic tasks. Input and output rates are unchanged, so for teams already on Fable 5 the move is mostly a re-point of the model id.
Who is it for?
teams running long agentic coding and research jobs
Frequently asked questions
- How much cheaper is Claude Fable 5.1 than Claude Fable 5?
- Claude Fable 5.1 keeps Fable 5's rates of $10 per million input tokens and $50 per million output tokens, but cache reads fall 75% to $0.25 per million. Anthropic puts the result at roughly 25% lower cost on typical workloads, and up to around 45% on complex coding and highly agentic tasks, where the same context is read back many times.
- How do I get access to Claude Mythos 5.1?
- Claude Mythos 5.1 is invitation-only through Project Glasswing, limited to US organizations in Anthropic's Cyber Verification Program and Life Sciences Verification Program. Anthropic says to contact your Anthropic, AWS or Google Cloud account team. Mythos 5.1 is the same underlying model as Claude Fable 5.1 with lighter safeguards, and shares its specifications and pricing.
- What breaks when moving from Claude Fable 5 to Fable 5.1?
- Anthropic lists three breaking changes for Claude Fable 5.1. Forced tool use now returns an error. Earlier Claude models cannot read its thinking blocks. Editing an earlier turn invalidates the thinking blocks in that conversation. Five other changes are additive and optional, including per-message effort, turn-scoped system messages, and readable progress updates between tool calls.
- What did Claude Mythos 5.1 do in the science experiments Anthropic reported?
- Anthropic reports that Claude Mythos 5.1 designed protein binders whose binding affinities were 10 times higher than the best entries submitted to Adaptyv Bio's protein design competitions on three targets, with a hit rate of nearly 50% across 12 targets. It also sped up seven open-source deep learning models by up to 2.5 times by writing custom GPU kernels.
Try it
Model id claude-fable-5-1 on the Claude API, Bedrock, Vertex AI and Microsoft Foundry