Anthropic launches Claude Fable 5.1 with lower costs and stronger coding performance

Anthropic has introduced Claude Fable 5.1 and Claude Mythos 5.1, a new model family aimed at coding, knowledge work, and scientific research. Anthropic says both versions use the same underlying model, but Mythos 5.1 has stricter access controls and is currently limited to trusted programs.

Fable 5.1 is generally available and is expected to cost about 25% less than Fable 5 for typical token-billed workloads. For highly agentic tasks, savings could reach approximately 45%, largely because Anthropic is cutting cache-read pricing—important for tools that repeatedly work against large, persistent context windows.

More capable, with fewer interruptions

Anthropic reports a 52.6% score for Fable 5.1 on Terminal-Bench-Science 0.1, versus 24.7% for Fable 5. On Terminal-Bench 4.0, Fable 5.1 scored 55.8%, while Mythos 5.1 reached 60.9%. Other reported results include 73.4% on CursorBench 3.2.0 and 65.0% on Humanity’s Last Exam when using tools.

For developers, the most practical change may be improved behavior on long-running tasks. Anthropic says the model is better at identifying root causes instead of applying superficial fixes. Investment firm Millennium reportedly used it to trace a rare crash to a bug in an external vendor library after years of unsuccessful investigation.

The company also says its cybersecurity safeguards now produce 60% fewer false positives. Fable 5.1 can help discover software vulnerabilities, but not develop exploits. In biology, Mythos 5.1 will be offered through an access program developed with the US government, with scientist enrollment expected soon.

Anthropic’s Enterprise Frontier Safeguards will provide customer-controlled data storage and zero-data-retention-level privacy, with enterprise rollout beginning later this fall. Until then, eligible customers can use Fable 5.1 with zero data retention.

Source: Anthropic News