Anthropic launches Claude Fable 5.1 and limited Mythos 5.1

Anthropic on 1 September 2026 introduced Claude Fable 5.1 and Claude Mythos 5.1, describing them as its most advanced models for coding and knowledge work. The company says the two names are the same underlying model with different safeguards: Fable 5.1 is generally available; Mythos 5.1 is limited to trusted-access programs for cybersecurity and life sciences.

Availability, pricing, and Claude Code

Fable 5.1 is available today on Anthropic’s platforms and through Amazon Web Services, Google Cloud, and Microsoft Azure (the product page also lists Microsoft Foundry). Developers can call claude-fable-5-1 on the Claude API. Anthropic’s Claude Fable product page says Pro, Max, Team, and Enterprise users can use it, and that it is built for long-running coding, code review, and unattended agent work.

Input and output prices match Fable 5: $10 per million input tokens and $50 per million output tokens. Cache reads now cost $0.25 per million tokens, which Anthropic says is 75% less than Fable 5. The company estimates that change cuts typical token-billed workloads by about 25%, and highly agentic work by up to about 45%, because cache reads dominate those jobs.

In Claude Code, Fable 5.1 defaults to High effort; on Claude.ai and Claude Cowork it defaults to Medium. Anthropic says Low or Medium effort on Fable 5.1 can match or beat Fable 5 at lower cost.

Benchmarks Anthropic published

Anthropic evaluated Fable 5.1 with production safeguards on. Selected scores from the launch post:

| Benchmark | Fable 5.1 | Fable 5 | Opus 5 | GPT-5.6 Sol | | --- | --- | --- | --- | --- | | Terminal-Bench-Science 0.1 | 52.6% | 24.7% | 29.0% | 22.4% | | Terminal-Bench 4.0 | 55.8% (60.9% Mythos 5.1) | 42.0% | 52.3% | 37.3% | | CursorBench 3.2.0 | 73.4% | 70.5% | 70.0% | 67.2% | | GDPval-AA v2 | 1853 | 1723 | 1824 | 1711 | | OSWorld 2.0 (strict) | 41.7% | 36.1% | 39.6% | — | | Humanity's Last Exam (with tools) | 65.0% | 63.8% | 63.6% | — | | AutomationBench | 31.4% | 17.1% | 26.9% | 19.6% |

Anthropic notes that when safeguards intervened, some tasks scored zero or fell back to Opus 4.8 (cyber) or Opus 5 (biology), which it says likely understates Fable scores. OSWorld 2.0 used the authors’ August 2026 task release, so those numbers are not comparable to older published OSWorld results.

Cognition’s Walden Yan said the company is moving Opus 5 traffic in Devin to Fable 5.1 on launch day after it “matched or edged out Fable 5” at lower cost per task. Cursor’s launch quotes on the same post say Fable 5.1 scored 73.4% at max effort on CursorBench 3.2 and was “especially skilled at verifying its own work.”

Safeguards, Mythos access, and enterprise data

Anthropic says Fable 5.1 can now be used to identify software vulnerabilities in source code, but not to develop exploits. Cyber safeguards still send penetration testing, exploit generation, and binary-based vulnerability scanning to Opus models. The company says Claude Code users should see about 60% fewer cyber-safeguard interventions per session versus the previous Fable 5 safeguards. Biology safeguards still route dual-use life-science R&D to Opus; Anthropic previously reported an 85% drop in false-positive biology fallbacks on elementary and medical questions.

Mythos 5.1 is for vetted US organizations via the Cyber Verification Program (Mythos-class access “in the near future”) and a Life Sciences Verification Program developed with the US government. Claude Security, which scans codebases and suggests patches for human review, is now powered by Mythos 5.1.

A companion post, Developing Enterprise Frontier Safeguards, describes Enterprise Frontier Safeguards (EFS): customer-controlled cloud storage (for example Amazon S3, Azure Blob Storage, or Google Cloud Storage), automated misuse detection, and customer-run human review by default. EFS is planned in phases later this fall on Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Google’s Agent Platform, and Microsoft Foundry. Eligible customers get zero data retention on Fable 5 and 5.1 until EFS is ready. Anthropic says it does not charge for EFS; cloud storage is billed by the customer’s cloud provider.

Fable 5.1 also includes stronger anti-distillation controls: new API accounts created from 1 September 2026 cannot manually edit Claude’s prior context in a multi-turn conversation while keeping the prior thinking transcript. Existing accounts are not currently affected.

Anthropic says Fable 5.1 and Mythos 5.1 outputs include the EU AI Act watermark required for models released after 2 August 2026, with a detection API in private preview for eligible organizations.

Source: Anthropic, Introducing Claude Fable 5.1 and Claude Mythos 5.1.