Qwen 3.5 35B A3B (Reasoning) — benchmark results
Provider: Alibaba. Released 2026-02-24. Access: Open.
Unified ELO 1612 ± 20, rank #503 of 1806 rated models, from 52 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA IFBench | 72.52 | Accuracy (%) | 89.5 |
| AA TAU-2 Bench | 89.18 | Accuracy (%) | 83.6 |
| AA GPQA Diamond | 84.55 | Accuracy (%) | 81.1 |
| AA Humanity's Last Exam | 21.04 | Accuracy (%) | 77.8 |
| AA Global-MMLU-Lite - Portuguese | 90.42 | Accuracy (%) | 77.7 |
| AA Global-MMLU-Lite - Italian | 90.42 | Accuracy (%) | 77.2 |
| AA Long Context Reasoning | 68.33 | Accuracy (%) | 77.2 |
| AA-LCR | 62.7 | Score (self-reported) | 76.8 |
| AA Global-MMLU-Lite - Bengali | 86.58 | Accuracy (%) | 76.3 |
| AA Global-MMLU-Lite - Burmese | 81.75 | Accuracy (%) | 75 |
| AA Global-MMLU-Lite - Chinese | 88.75 | Accuracy (%) | 74.1 |
| AA Global-MMLU-Lite - Yoruba | 69 | Accuracy (%) | 73.9 |
Interactive version: theaggregate.ai/model?slug=qwen-3-5-35b-a3b-reasoning · How It Works · Data refreshed daily, snapshot 2026-08-07.