Qwen 3.5 397B A17B (Reasoning): benchmark results
Provider: Alibaba. Released 2026-02-16. Access: Open.
Unified ELO 1653 ± 1, rank #317 of 3081 rated models, from 148 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| t2-bench | 87.91 | Score (%) | 100 |
| AA Global-MMLU-Lite - Italian | 93.33 | Accuracy (%) | 97.3 |
| AA IFBench | 78.78 | Accuracy (%) | 97.1 |
| AA Global-MMLU-Lite - Spanish | 93.17 | Accuracy (%) | 95.5 |
| Tau2-Bench Telecom | 95.61 | Success Rate (%) | 95.4 |
| AA TAU-2 Bench | 95.61 | Accuracy (%) | 95.3 |
| AA Global-MMLU-Lite - English | 94 | Accuracy (%) | 95 |
| AA Global-MMLU-Lite - French | 92.75 | Accuracy (%) | 93.6 |
| AA Global-MMLU-Lite - Korean | 90.25 | Accuracy (%) | 91.8 |
| AA Global-MMLU-Lite - Portuguese | 92.25 | Accuracy (%) | 91.8 |
| AA Global-MMLU-Lite - Yoruba | 80.58 | Accuracy (%) | 90.9 |
| AA Global-MMLU-Lite - Chinese | 91.08 | Accuracy (%) | 89.5 |
Interactive version: theaggregate.ai/model?slug=qwen-3-5-397b-a17b-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-21.