AI modelsNovember 09, 2026
-4
❄️
Claude Fable 5.2 cracks Fermat in 30 hours
Anthropic has introduced two new models, Fable 5.2 and Mythos 5.2, which improve on coding tasks while offering faster and cheaper inference. The release comes amid ongoing discussion about rapid self-improvement in AI systems.
#Anthropic#Claude#LLM#Model Release#Benchmarks

🔥 What happened
Anthropic released Claude Fable 5.2 and Mythos 5.2. The model formalized Fermat's Last Theorem in 30 hours, designed three working electronic devices from scratch, and solved two open problems in black hole physics.
💡 Why it matters
Fable 5.2 is 30% faster and cheaper than its predecessor: $8 per million input tokens, $40 per million output tokens. Fast mode doubles the price but speeds up token generation by 2.5x. On FrontierCode 2.0, it matches Opus 5.5 at roughly half the cost per task.
⚡ Our take
Anthropic is positioning Fable as the "Advisor" for the hardest problems – a smart move against OpenAI and Google. If you don't jump on the advisor bandwagon now, you'll pay for it later.
The title, summary and analysis of this item were produced automatically by an AI system and have not been editorially reviewed. They may contain errors, bias or omissions — when in doubt, read the linked original source.
Deep Dives & Similar Intelligence
Anthropic opens Claude Skills to all
Tools & ProjectsNovember 09, 2026

Anthropic Lets Claude Dream Between Sessions
ResearchNovember 06, 2026

Anthropic Nears the Self-Improvement Threshold
Ethics & SecurityNovember 03, 2026

AI Models Smuggle Secret Messages in Images
ResearchNovember 08, 2026

EnvOpt: Your Agent's Environment Is the New Model
ResearchNovember 05, 2026