LaunchAI ModelsNovember 10, 2026

Anthropic ships Claude Fable 5.2 and Mythos 5.2, retaking top LLM spot

Read original source →latent.space

Fable 5.2 cuts per-token pricing 20% to $8/M input and $40/M output, but SemiAnalysis's GDPx benchmark shows cost per task rising to $88.15 from $67.16 as the model runs far more test cases and subagents. It took first place on GDPx at 16.4%, displacing Gemini 4 Carbon after 10 days.

1 source

More stories today

Open the live feed