AI Radar

Noteworthy AI conversations and writing, summarized from publisher notes.

Claude Fable/Mythos 5.1: new SOTA model, 75% cache price cut but 70% more output tokens

· Latent Space

Anthropic launched Claude Fable 5.1 and Claude Mythos 5.1, positioning them as the world's most advanced models for coding and knowledge work, particularly for autonomous, multi-step tasks. The models retain previous pricing for input/output/cache write but feature a 75% cache read price cut to $0.25/MTok, benefiting long sessions and agentic workloads. Fable 5.1 has a 1M token context window and supports image+text inputs, topping Artificial Analysis's Intelligence Index at 66. However, it is 20% more expensive per task than Fable 5 due to using approximately 1.7 times more output tokens. Community analysis suggests Fable and Mythos 5.1 may share identical weights, with differences primarily in safety/routing behavior, including server-side fallbacks for ~4% of output tokens in AA's evaluations.

AI Radar wrote this summary from the publisher’s notes with AI, not from a transcript. How it works.