reasoning benchmark
TAU3-Bench
TAU3-Bench is a benchmark for evaluating general-purpose agent capabilities, testing models on multi-turn interactions with simulated user models, retrieval, and complex decision-making scenarios.
Updated Aug 7, 2026
Published models5
Registry coverage5
MetricScore
EvidenceB
TAU3-Bench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | XI | 72.9% | 100.0% | 5 | C | |
| 2 | AC | 70.7% | 75.0% | 5 | C | |
| 3 | ZA | 70.6% | 50.0% | 5 | C | |
| 4 | AC | 67.2% | 25.0% | 5 | C | |
| 5 | NV | 22.6% | 0.0% | 5 | C |
Score distribution
Top published rows on the benchmark's original scale.
TAU3-Bench highlights
The top published results on this benchmark's own scale.
What is TAU3-Bench?
Definition and scoring fields from the benchmark registry.
TAU3-Bench is a benchmark for evaluating general-purpose agent capabilities, testing models on multi-turn interactions with simulated user models, retrieval, and complex decision-making scenarios.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- TAU3-Bench
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- tau3-bench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about TAU3-Bench.
Which model scores highest on TAU3-Bench?
MiMo-V2.5-Pro is currently ranked first with 72.9%.
What does TAU3-Bench measure?
TAU3-Bench is a benchmark for evaluating general-purpose agent capabilities, testing models on multi-turn interactions with simulated user models, retrieval, and complex decision-making scenarios.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
5 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.