Claude Fable 5.2 and Mythos 5.2

708 points · 877 comments on HN · read original →

Points and comments are a snapshot, not live.

Anthropic releases Claude Fable 5.2 and Mythos 5.2 with faster speeds and lower pricing.

Claude Fable 5.2 and Mythos 5.2 show improvements in coding, reasoning, and vision. They are 30% faster than version 5.1 and cost $8 per million input tokens and $40 per million output tokens. Fast mode offers up to 2.5x speed at $16/$80. Fable has cybersecurity and biology safeguards; Mythos lacks those safeguards and is available via verification programs. Benchmarks show Fable 5.2 outperforms Opus 5.5 on most evaluations including agentic coding (FrontierCode 49.1%, FrontierSWE 69.9%), cybersecurity, and scientific research. The model formalized Fermat's Last Theorem in 30 hours using 1,000 agents, and solved two open physics problems about black holes.

What commenters are saying

Commenters question the naming, noting leaks suggested "Fable 5.5" and confusion that 5.2 beats Opus 5.5 on benchmarks. Some speculate the company is saving 5.5 for a larger update or to counter upcoming GPT-6 Astra 6.2. Several users criticize Claude's writing style as dense and cryptic, though some find Opus 5.5 improved. Pricing at $200/month for unlimited use is defended as cost-effective by some, while others say it remains too expensive. The generous reset terms for long-term subscribers receive positive attention.