Anthropic Launches Mythos 5, Fable 5 Record-Breaking Models

Anthropic has introduced Claude Mythos 5 and Claude Fable 5, two large language models that outperform competitors across multiple benchmarks, including an 80.3% score on SWE-Bench Pro. Fable 5 is publicly available with strict guardrails blocking high-risk prompts, while Mythos 5 has relaxed restrictions and limited access managed with the U.S. government. Mythos 5 can generate novel scientific hypotheses, discovering drug protein targets and new E. coli protein data in internal tests. Both models are priced at $10 per million input tokens and $50 per million output tokens, less than half the cost of their predecessor.