Source time Edition date
Anthropic launches new Claude Fable and Mythos 5.2 models
Anthropic推出Claude Fable与Mythos 5.2新模型
Anthropic's claim that Claude Fable 5.2 cracked two open theoretical physics problems is striking, but physicists must independently verify the reasoning before anyone trusts a model's math.
Judgment
Physicists and engineering leads deciding whether to trial Claude Fable 5.2 must weigh Anthropic's claimed black-hole problem results against the cost of independent verification, since only vetted organizations get Mythos 5.2 access.
Basis
Anthropic says Claude Fable 5.2 and Mythos 5.2 improve coding, reasoning, and vision and generate outputs faster than 5.1. In prerelease testing, Anthropic claims the models helped engineers complete software migrations, designed three working electronic devices, solved two open problems in the physics of black holes, made progress on robotics evaluations, and formalized Fermat's Last Theorem in 30 hours. Fable 5.2 launches today; Mythos 5.2 is limited to vetted organizations. The mechanism is agent-team orchestration and long-running project support, but the claims rest on Anthropic's own testing.
Unclear The source does not say whether the physics solutions have been peer-reviewed, what the problems were, or how verification would be conducted.
Next Watch for independent physicists or a peer-reviewed venue to confirm or dispute the claimed black-hole solutions before treating them as validated.