T1-Bench - Pass^3: leaderboard

Metric: Pass^3 (%): share of conversations solved in all three trials, where a conversation passes when every ground-truth tool call and parameter is reproduced and every ground-truth tool output is matched; T1-Bench: 525 multi-turn customer-service conversations over 25 domain settings (11 single domains and 14 combinations of 2 to 11 domains, 76 tools), a Gemma4-26B-A4B-it user simulator, three trials per conversation at temperature 1.0; averages weighted by conversations per setting; higher is better. Source: arxiv.org. Saturation forecast: Around 2035. 12 models tracked.

Top models

#ModelScore
1Gemma 4 31B (IT)41.14
2Gemma 4 26B A4B (IT)37.71
3gemma-4-E4B-it35.62
4Claude Opus 4.634.48
5Claude Haiku 4.534.29
6Claude Sonnet 4.533.71
7GPT-OSS-120B (High)29.33
8Gemma 4 E2B (IT)28.38
9GPT-5.4 (Low)20.95
10GPT-OSS-20B (High)8.57

Interactive version: theaggregate.ai/benchmark?slug=t1-bench-pass-pow-3 · How It Works · Data refreshed daily, snapshot 2026-09-29.