Anthropic has released Claude Fable 5.1 and Mythos 5.1, achieving top performance across eight benchmarks and reducing prices by up to 45% for certain tasks. Fable 5.1 demonstrates improved multi-step reasoning and better adherence to stylistic prompts, while Mythos 5.1 showcases significant advancements in scientific research and coding, including a 10x improvement in protein design affinity and a 2.5x speedup in computational biology tasks. The release also introduces a new anti-distillation mechanism to prevent manipulation of conversation history and context. AI
IMPACT Sets new SOTA on scientific and coding benchmarks, potentially lowering costs for complex AI tasks and improving research capabilities.
RANK_REASON Frontier-lab model release with system card and benchmark results. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Adaptyv Bio
- Anthropic
- Claude
- EnVision
- European Space Agency
- Fable 5.1
- Flix Rieseberg
- GPT 5.6 "Sol"
- Mythos 5.1
- Nasa
- Veritas
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →