Skip to content
Bencher

TestsBenchmarks · Tool use

τ³-bench Retail

τ³-bench retail domain (Sierra).

% solvedHigher is betterThe test's websiteOfficial page ↗

The best 2 · 2 models tested

Top 2 · best setting per model · 2 models, 2 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported
τ³-bench Retail leaderboard
Thinking levelSetting
1Mistral Medium 3.5Mistral AI76.1%Highmaximum reasoning settings (reasoning_effort=high)Maker's own figureVendor-reportedMistral AI ↗—22 May 2026
2Mistral Small 4Mistral AI67.8%Defaultreasoning setting not statedMaker's own figureVendor-reportedMistral AI ↗—22 May 2026