Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026. According to Anthropic’s announcement, “Claude Fable 5.1 and Claude Mythos 5.1 are the same model, but with different levels of safeguards.” Fable 5.1 is generally available; Mythos 5.1 is limited to vetted organizations in cybersecurity and the life sciences.
On this page
One model, two access levels
Anthropic says “Claude Fable 5.1 is available today on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure.” The API model ID is claude-fable-5-1.
Mythos 5.1 is the restricted-access version. Anthropic says it “is available to vetted cyberdefenders and life scientists. Currently, it is only available to a set of US organizations.” Access goes through two programs: the Cyber Verification Program and the Life Sciences Verification Program.
Safeguards that fire less often
Most of the announcement’s safety changes concern how often Fable 5.1’s safeguards step in on legitimate work.
- Cybersecurity: Anthropic says it updated its cyber safeguards “to be more precise” and now allows Fable 5.1 to be used for identifying software vulnerabilities, described as “the kind of defensive work that improves software security.” As a result, the company says “Claude Code users can expect an average of around 60% fewer interventions per session from our cyber safeguards, relative to the previous safeguards on Fable 5.”
- Biology: Anthropic says its latest biology safeguards for Fable 5.1 and Fable 5 “fire 85% less often for benign requests related to elementary biology and medical questions,” compared with those that launched with Fable 5.
Anthropic also says new API accounts cannot “manually edit Claude’s prior context in a multi-turn conversation while preserving the transcript,” a measure aimed at distillation attacks. In its Claude Opus 5.5 announcement, Anthropic calls this “preserved thinking, the anti-distillation safeguard we introduced with Fable 5.1.”
Benchmark results
All scores below were published by Anthropic for Claude Fable 5.1.
| Benchmark | Claude Fable 5.1 | Anthropic’s category / setting |
|---|---|---|
| Terminal-Bench 4.0 | 55.8% | Agentic coding |
| Terminal-Bench-Science 0.1 | 52.6% | Agentic scientific research |
| Humanity’s Last Exam | 60.9% | No tools |
| CursorBench 3.2.0 | 73.4% | Agentic coding |
Anthropic also reports 65.0% on Humanity’s Last Exam with tools. Lab-published scores for Fable 5.1 and other models are tracked on our benchmark leaderboard.
What it changes
Anthropic calls the pair “the world’s most advanced models for coding and knowledge work.” The split release puts one model in front of all customers under standard safeguards, while vetted US security and life-science organizations use the same model under a different safeguard level. For Claude Code users doing defensive security work on Fable 5.1, the practical change is fewer interventions from the cyber safeguards and permission to use the model for vulnerability discovery.
Both models are listed on our AI model release timeline.
