Anthropic's Claude Fable 5.2 Beats GPT-6 Cyber in Multi-Stage Cyberattack Benchmark

Anthropic on Monday released Claude Fable 5.2 and Claude Mythos 5.2 models, which it said exhibit improvements in coding, reasoning, and vision. During pre-release training, the large language models (LLMs) helped its engineers "complete ot