Qwen 3.7 Flash: benchmark results
Alibaba's Qwen 3.7 Flash (July 2026), a native vision-language model for search-agent and code-interpreter workflows, with a 991K-token context and reasoning on by default but toggleable per request. Provider: Alibaba. Released 2026-07-25. Access: API.
Aggregate Rating 1722 ± 24, rank #156 of 1635 rated models, from 49 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MERA v2 - SAGE | 68.2 | Score (%) | 87.2 |
| AI for Education Pedagogy - Science | 91.26 | Accuracy (%) | 84.2 |
| ClashBench - Deliberate Interference Rate (Claude Code) | 41.4 | DIR (%) | 83.3 |
| Tinybird AI SQL Benchmark - Exactness | 54.09 | Result exactness vs human reference queries (0-100) | 82.1 |
| MERA v2 - GorillaHard | 50.6 | Score (%) | 78.2 |
| AI Chess Leaderboard (Reasoning) | 943 | Elo | 72.6 |
| Tinybird AI SQL Benchmark - Success Rate | 100 | Questions answered with a valid query within 3 retries (%) | 71.7 |
| AI for Education Pedagogy - Primary | 88.73 | Accuracy (%) | 70.1 |
| AI for Education Pedagogy - Secondary | 84.75 | Accuracy (%) | 70.1 |
| AI for Education Pedagogy - Social studies | 83.64 | Accuracy (%) | 70.1 |
| OTIS Mock AIME 2024-25 | 86.67 | Accuracy (%) | 70 |
| AI for Education Pedagogy - Maths | 84.92 | Accuracy (%) | 69.9 |
Interactive version: theaggregate.ai/model?slug=qwen-3-7-flash · How It Works · Data refreshed daily, snapshot 2026-10-15.