reasoning benchmark
Tau2 Retail Leaderboard
Tau2 Retail evaluates conversational AI agents in retail customer service scenarios within a dual-control environment where both the agent and user can interact with tools.
Updated Sep 5, 2026
Models27
Model coverage27
MetricScore
EvidenceB
Tau2 Retail Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score91.90% | Percentile100.00% | Participants27 | EvidenceC | Evaluated |
| Rank02 | ModelAN | Score91.70% | Percentile96.15% | Participants27 | EvidenceC | Evaluated |
| Rank03 | ModelAN | Score88.90% | Percentile92.31% | Participants27 | EvidenceC | Evaluated |
| Rank04 | ModelME | Score88.60% | Percentile88.46% | Participants27 | EvidenceC | Evaluated |
| Rank05 | ModelAN | Score83.20% | Percentile84.62% | Participants27 | EvidenceC | Evaluated |
| Rank06 | ModelOP | Score82.00% | Percentile80.77% | Participants27 | EvidenceC | Evaluated |
| Rank07 | ModelOP | Score81.10% | Percentile76.92% | Participants27 | EvidenceC | Evaluated |
| Rank08 | ModelOP | Score80.20% | Percentile73.08% | Participants27 | EvidenceC | Evaluated |
| Rank09 | ModelAM Nova 2 OmniAmazon | Score78.30% | Percentile69.23% | Participants27 | EvidenceC | Evaluated |
| Rank10 | ModelLA | Score77.90% | Percentile65.38% | Participants27 | EvidenceC | Evaluated |
| Rank11 | ModelOP | Score77.90% | Percentile61.54% | Participants27 | EvidenceC | Evaluated |
| Rank12 | ModelOP | Score77.90% | Percentile57.69% | Participants27 | EvidenceC | Evaluated |
| Rank13 | ModelOP | Score77.90% | Percentile53.85% | Participants27 | EvidenceC | Evaluated |
| Rank14 | ModelAMNova 2 ProAmazon | Score77.70% | Percentile50.00% | Participants27 | EvidenceC | Evaluated |
| Rank15 | ModelAMNova 2 LiteAmazon | Score76.50% | Percentile46.15% | Participants27 | EvidenceC | Evaluated |
| Rank16 | ModelME | Score73.10% | Percentile42.31% | Participants27 | EvidenceC | Evaluated |
| Rank17 | ModelAC | Score71.90% | Percentile38.46% | Participants27 | EvidenceC | Evaluated |
| Rank18 | ModelME | Score71.50% | Percentile34.62% | Participants27 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score71.30% | Percentile30.77% | Participants27 | EvidenceC | Evaluated |
| Rank20 | ModelME | Score71.27% | Percentile26.92% | Participants27 | EvidenceC | Evaluated |
| Rank21 | ModelMA | Score70.60% | Percentile23.08% | Participants27 | EvidenceC | Evaluated |
| Rank22 | ModelMA | Score70.60% | Percentile19.23% | Participants27 | EvidenceC | Evaluated |
| Rank23 | ModelAC | Score67.80% | Percentile15.38% | Participants27 | EvidenceC | Evaluated |
| Rank24 | ModelOP | Score63.40% | Percentile11.54% | Participants27 | EvidenceC | Evaluated |
| Rank25 | ModelNV | Score62.83% | Percentile7.69% | Participants27 | EvidenceC | Evaluated |
| Rank26 | ModelAC | Score57.30% | Percentile3.85% | Participants27 | EvidenceC | Evaluated |
| Rank27 | ModelNV | Score56.90% | Percentile0.00% | Participants27 | EvidenceC | Evaluated |
Tau2 Retail Highlights
The leading models and scores on this benchmark.
Tau2 Retail Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Tau2 Retail
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is Tau2 Retail?
What Tau2 Retail measures and how its scores work.
Tau2 Retail is a benchmark for conversational AI agents in retail customer support contexts.
It measures tool-agent-user interaction, rule adherence, and task consistency, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Tau2 Retail
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Tau2 Retail.
Which model scores highest on Tau2 Retail?
Claude Opus 4.6 is currently ranked first with 91.90%.
What are the top three models on Tau2 Retail?
The current leaders are Claude Opus 4.6 (91.90%), Claude Sonnet 4.6 (91.70%), and Claude Opus 4.5 (88.90%).
Which Tau2 Retail model has the lowest official input price?
Nemotron 3 Super (120B A12B) has the lowest matched official input price at $0.20 input / $0.80 output per 1M tokens.
Which models are fastest among Tau2 Retail results?
The fastest matched records are GPT-4o (132.00 tok/s via OpenAI), Claude Opus 4.6 (123.15 tok/s via Anthropic), and Claude Haiku 4.5 (100.00 tok/s via Vertex AI).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.