Sierra Launches τ^τ-bench, Humans Far Outpace Top AI Models

Sierra released τ^τ-bench, a new benchmark for multi-turn tool calling, where humans scored 82.2% versus Claude Opus 5's 23.9%. Kimi K3's inconsistent scores two days apart also sparked skepticism.

2026-09-09 ~ 2026-09-09 · 2 related posts

• Sierra's New Tool-Calling Bench τ^τ-bench: Kimi K3 Scores Differ Two Days Apart, Banking Sub-bench Hardest — zainhas · 2026-09-09
• Sierra's τ^τ-bench: Humans Hit 82.2% While Best Model Claude Opus 5 Manages Only 23.9% — zainhas · 2026-09-09