Key facts
- Anthropic released Claude Fable 5.1 and Mythos 5.1, featuring performance upgrades and cost reductions.
- Fable 5.1 significantly improved benchmark scores, particularly in scientific and coding tasks.
- Mythos 5.1, with lighter safety filters, is available to vetted professionals in cybersecurity and life sciences.
- Cache read costs have been reduced by 75%, leading to lower overall workload expenses.
- Access to Fable 5.1 is tiered based on subscription plans, with some including usage and others relying on credits.
Anthropic has launched its latest AI models, Claude Fable 5.1 and Claude Mythos 5.1, demonstrating significant performance gains over their predecessors. Fable 5.1 achieved a 52.6% pass rate on the Terminal-Bench-Science 0.1 benchmark, more than doubling the 24.7% score of Fable 5, and scored 55.8% on Terminal-Bench 4.0, an improvement from Fable 5's 42.0%. Mythos 5.1, a variant with less restrictive safety filters, achieved a 60.9% score on the latter benchmark.
These new models also bring cost efficiencies, with cache reads now costing 75% less. This reduction translates to approximately 25% lower costs for typical workloads and up to 45% savings for highly agentic workloads, while base pricing remains unchanged at $10 per million input tokens and $50 per million output tokens.
Access to Fable 5.1 is tiered. It is included for up to 50% of weekly limits on Max plans and premium Team or Enterprise seats. For Pro plans and standard Team seats, Fable 5.1 is accessed via usage credits billed at the API rate, as it is not part of those plans' included weekly limits.
Fable 5.1 and Mythos 5.1 are essentially the same model with different safety configurations. Mythos 5.1 is restricted to cybersecurity and life-sciences professionals through Anthropic's verification programs. The models are available through the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.