LaunchAI ModelsJuly 2, 2026

Anthropic relaunches Claude Fable 5 with safety guardrails

Anthropic re-enabled Claude Fable 5 with visible safety fallbacks, routing some requests to Opus 4.8, and reset rate limits. Cursor, Devin, and Perplexity quickly added support, though Cursor noted it's the most expensive model per task.

3 sources

Anthropic relaunches Claude Fable 5 with safety guardrails — AIBriefs