Qwen 3 Max (Thinking): benchmark results
Current thinking snapshot of Qwen 3 Max. Provider: Alibaba. Released 2026-01-27. Access: API.
Aggregate Rating 1664 ± 9, rank #477 of 2168 rated models, from 181 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AgentNoiseBench (tau2-bench, User Noise) | 78 | Success on a stratified 25 percent sample of tau2-bench task | 100 |
| GISA - List F1 | 66.51 | F1 (%) of list items regardless of order on the 48 list-type | 100 |
| GISA - List Order Score | 64.08 | Order score (%) on the 48 list-type queries: twice the match | 100 |
| GISA - Table Item F1 | 66.86 | Item-level F1 (%) between predicted and gold table cells on | 100 |
| LLM Stats (C-Eval) | 93.7 | Score (%) | 100 |
| Tau-Bench Telecom | 98.2 | Pass@1 (%) | 100 |
| AI Chess Leaderboard (Reasoning) | 1800 | Elo | 98.5 |
| BenchTable - Tech | 85.5 | Weighted Score (%) | 94.8 |
| Tinybird AI SQL Benchmark - First-Attempt Success Rate | 100 | Questions answered with a valid query on the first attempt ( | 94 |
| BenchTable - STEM | 82.9 | Weighted Score (%) | 93.6 |
| GISA - Set EM | 30 | Exact match (%) on the 50 set-type queries (an unordered col | 93.3 |
| GISA - Table EM | 10.67 | Exact match (%) on the 253 table-type queries (a table with | 93.3 |
Interactive version: theaggregate.ai/model?slug=qwen-3-max-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.