Claude Fable 5.2 Arrives in Factory

This is a strong model-and-harness pairing: Fable 5.2 supplies deeper reasoning, while Factory’s workflow emphasizes testing, diagnosis, and delivery. The main tradeoff is cost, making it best suited for difficult tasks where correctness matters more than throughput.

• –Factory reports stronger performance on diagnostics, not just code that runs
• –A 1M-token context window supports large repositories and legacy migrations
• –Adaptive thinking and configurable effort let teams trade latency and cost for depth
• –At $8 per million input tokens and $40 per million output tokens, it is a premium escalation model
• –Developers should validate the claims against their own repositories and migration benchmarks