Skip to content

Benchmark

tau2-bench v0.1 retail pass^4

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
6
Scores recorded
7
Sources
1
IQ weight
1.03

Score distribution

52.8%
Mean
57%
Median
7.7%
Std dev
46.8%
10th percentile
58.8%
90th percentile
36.9%
Min
36.9%59.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.03 for chat models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Sonnet 3.7submission-claude-3-7-sonnet_anthropic_2024-06-20, usersim-gpt-4.1-2025-04-14· Sierra tau-bench Leaderboard data docs59.7%
2GPT 5submission-gpt-5_sierra_2025-08-09, usersim-gpt-4.1-2025-04-14· Sierra tau-bench Leaderboard data docs58.8%
3Qwen3 Maxsubmission-qwen3-max_qwen_2026-01-23, usersim-gpt-4.1-2025-04-14· Sierra tau-bench Leaderboard data docs57.9%
4Qwen3 Maxsubmission-qwen3-max_qwen_2024_09_23· Sierra tau-bench Leaderboard data docs57%
5GPT 4.1submission-gpt-4-1_openai_2024-06-20, usersim-gpt-4.1-2025-04-14· Sierra tau-bench Leaderboard data docs52.3%
6o4 Minisubmission-o4-mini_openai_2024-06-20, usersim-gpt-4.1-2025-04-14· Sierra tau-bench Leaderboard data docs46.8%
7GPT-4.1 minisubmission-gpt-4-1-mini_openai_2024-06-20, usersim-gpt-4.1-2025-04-14· Sierra tau-bench Leaderboard data docs36.9%

Data credits: Sierra Research tau2-bench leaderboard submissions (github.com/sierra-research/tau2-bench), MIT (licence)