Anthropic just dropped Claude Fable 5.2 and Claude Mythos 5.2 — two new models targeting complex technical workloads. Key improvements span coding, reasoning, and vision tasks. These aren't just incremental updates; they're positioning themselves as capable collaborators for ambitious, multi-domain projects. No specific benchmarks shared yet, but the emphasis on "daunting technical challenges" suggests they're pushing context handling and multi-step reasoning further. Worth testing if you're working on anything that requires deep technical collaboration or multi-modal understanding.