Qwen 3.5 9B (Thinking): benchmark results
Provider: Alibaba. Released 2026-03-02. Access: Open.
Aggregate Rating 1579 ± 8, rank #784 of 2170 rated models, from 242 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CanLegalRAGBench (Oracle Retrieval) - Answer Accuracy | 76.46 | Claim accuracy (%; share of the answer's atomic claims entai | 100 |
| JetBrains Mellum2.1 Card - AIME 2025-2026 | 86.7 | Exact match (%) | 100 |
| JetBrains Mellum2.1 Card - GPQA Diamond | 77.8 | Accuracy (%) | 100 |
| JetBrains Mellum2.1 Card - GSM-Plus | 91.4 | Exact match (%) | 100 |
| JetBrains Mellum2.1 Card - HarmBench | 6.6 | Harmful rate (%) | 100 |
| JetBrains Mellum2.1 Card - IFEval | 92.4 | Accuracy (%) | 100 |
| JetBrains Mellum2.1 Card - MMLU-Redux | 89.5 | Accuracy (%) | 100 |
| JetBrains Mellum2.1 Card - MixEval-Hard | 50.3 | Accuracy (%) | 100 |
| JetBrains Mellum2.1 Card - SWE-bench Pro (Pi harness) | 38 | Resolved (%) | 100 |
| JetBrains Mellum2.1 Card - SWE-bench Verified (Pi harness) | 50 | Resolved (%) | 100 |
| JetBrains Mellum2.1 Card - Terminal-Bench 2.1 (Pi harness) | 21.7 | Resolved (%) | 100 |
| JetBrains Mellum2.1 Card - ToolHop | 52 | Accuracy (%) | 100 |
Interactive version: theaggregate.ai/model?slug=qwen-3-5-9b-thinking · How It Works · Data refreshed daily, snapshot 2026-11-05.