Anthropic Claude Fable 5.2 launches on B.AI, with long-context and Agent capabilities upgraded again

Anthropic's Claude Fable 5.2 launched on the B.AI platform on November 10, opening simultaneously via API and Web Chat for complex reasoning, software engineering, and long-running agent workflows. The model runs with adaptive thinking always enabled and offers a 1M-token context window with up to 128K output tokens, and it topped the Artificial Analysis Intelligence Index v5 with a score of 65 at its highest configuration versus a peer median of 40. B.AI priced standard usage at $8.00 per million input tokens and $40.00 per million output tokens, with prompt caching supported and generation speeds up to 30% faster than Fable 5.1. The release matters because it pairs frontier-level benchmark performance with faster throughput and tiered pricing aimed at both heavy inference and high-frequency calls. Next, watch adoption across B.AI's API and chat surfaces and whether the 1M-token window and agent capabilities draw developers building long-horizon workflows.