Tau2 Retail
τ²-bench retail domain evaluates conversational AI agents in customer service scenarios within a dual-control environment where both agent and user can interact with tools. Tests tool-agent-user interaction, rule adherence, and task consistency in retail customer support contexts.
Claude Opus 4.6 from Anthropic currently leads the Tau2 Retail leaderboard with a score of 0.919 across 23 evaluated AI models.
Claude Opus 4.6 leads with 91.9%, followed by Claude Sonnet 4.6 at 91.7% and Claude Opus 4.5 at 88.9%.
Progress Over Time
Interactive timeline showing model performance evolution on Tau2 Retail
Tau2 Retail Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 2 | Anthropic | — | 200K | $3.00 / $15.00 | ||
| 3 | Anthropic | — | — | — | ||
| 4 |  Meituan | 560B | — | — | ||
| 5 | Anthropic | — | 200K | $1.00 / $5.00 | ||
| 6 |  OpenAI | — | 400K | $1.75 / $14.00 | ||
| 7 | OpenAI | — | — | — | ||
| 8 | OpenAI | — | — | — | ||
| 9 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 9 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 9 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 12 | Meituan | 69B | 256K | $0.10 / $0.40 | ||
| 13 |  Alibaba Cloud / Qwen Team | 235B | — | — | ||
| 14 | Meituan | 560B | — | — | ||
| 15 | Alibaba Cloud / Qwen Team | 235B | — | — | ||
| 16 | Meituan | 560B | 128K | $0.30 / $1.20 | ||
| 17 |  Moonshot AI | 1.0T | — | — | ||
| 17 | Moonshot AI | 1.0T | — | — | ||
| 19 | Alibaba Cloud / Qwen Team | 80B | — | — | ||
| 20 | OpenAI | — | 128K | $2.50 / $10.00 | ||
| 21 | 120B | — | — | |||
| 22 | Alibaba Cloud / Qwen Team | 80B | — | — | ||
| 23 | 32B | 262K | $0.06 / $0.24 |
FAQ
Common questions about Tau2 Retail.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions