Anthropic releases new Mythos and Fable models amid RSI concerns

Anthropic has released new versions of its advanced models, Mythos 5.2 and Fable 5.2, which show improvements in coding, reasoning, and vision, and communicate more clearly with users. Mythos 5.2, in particular, shows stronger covert capabilities than prior models, but Anthropic maintains it does not have the ability to act covertly against its users. The new Fable model is now cheaper than its main competitor, GPT-6 Astra, and is compatible with zero data retention policies. Anthropic has also addressed concerns about recursive self-improvement (RSI) by releasing the new models, but acknowledges ongoing challenges in testing model capabilities.