Claude Fable 5

Claude Fable 5 is Anthropic’s flagship model as of mid-2026, achieving an unprecedented 80.3% on SWE-Bench Pro — the highest score ever recorded on that benchmark at time of launch. It is purpose-built for long-horizon autonomous coding runs, meaning it can hold and act on multi-thousand-file codebases, multi-hour tasks, and deeply interconnected dependencies without losing context or coherence.