Fable 5.1 tops SWE-Bench, matching prior perf at 50% cost
ajratner · x · 2026-09-02
Claude Fable 5.1 now tops the Senior SWE-Bench leaderboard, edging out Fable 5 on the pass^3 tie-breaker by shipping quality code more consistently. Its best effort tier is Medium, and combined with cheaper cache reads, it matches Fable 5's performance at about 50% of the cost.
Related event: Fable 5.1 Tops Benchmarks but Costs More Per Task(16 posts)→
More from coding & agent
- Testing Fable 5.1 on optimizing Git implementation speed — samgoodwin89 · 2026-09-02
- Post: When your AI workflow succeeds but the result is wrong, how much trace do you inspect? — Sensitive-Parsnip-12 · 2026-09-02
- Querying databases through an MCP server in production instead of a GUI? — Green_Competition_21 · 2026-09-02
- Stripe Exec: AI Agents Dramatically Speed Up Product Testing Cycles — jeff_weinstein · 2026-09-02
- Google Cloud Run integrates Gemini Agent Platform with Identity and Registry — steren · 2026-09-02
- Grok picks 12 influential agent skill repositories out of 300+ studied — garrytan · 2026-09-02