Anthropic releases Claude Fable 5.1, doubling predecessor on key benchmarks
In brief
- Fable 5.1 scores 52.6% on Terminal-Bench-Science 0.1, more than double Fable 5's 24.7%
- Model beats Opus 5 on every published benchmark, costs twice as much per token
- Mythos 5.1 is the same model with stricter filters, restricted to vetted security professionals
Benchmark gains and pricing trade-offs
Fable 5.1 scored 55.8% on Terminal-Bench 4.0, up from Fable 5's 42.0% and ahead of Opus 5's 52.3%. On Humanity's Last Exam, Fable 5.1 reached 60.9% without external tools and 65.0% with them. The gains come at a cost: Fable 5.1 costs $10 input and $50 output per token, twice as much as Opus 5's $5 and $25.
Anthropic claims the new models are the world's most advanced for coding and knowledge work. It now beats Opus 5 on every benchmark Anthropic published. For teams already committed to Opus 5's price point, the performance gains may not justify the doubling of per-token cost. For others, the benchmark lead is substantial.
The Mythos variant and export control aftermath
Mythos 5.1 is the same underlying model with different safety filters bolted on, restricted to vetted cybersecurity and life-sciences professionals. Mythos 5.1 scored 60.9% on Terminal-Bench 4.0 with its lighter cybersecurity filters intact.
The release comes three months after Fable 5 was pulled offline for 18 days due to export control shutdown. Fable 5 launched on June 9, and Opus 5 launched in July at half of Fable 5's per-token price. The new models are available now on the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.
