Claude Fable 5.2 Refuses Less Often Than 5.1, but Its Gains Vary in Testing

Anthropic's Claude Fable 5.2 promises fewer refusals and lower token prices than Fable 5.1. The New Stack's hands-on comparison found those changes played out very differently across its tests. The excerpt is truncated, so the full benchmark results and the details of the login fix are not available here.

Why it matters: Teams choosing or upgrading between Claude model versions need to know that vendor claims on refusals and token pricing may not hold uniformly across real workloads.