Qwen 3.5 35B A3B Base: benchmark results
Provider: Alibaba. Released 2026-02-24. Access: Open.
Unified ELO 1684 ± 22, rank #264 of 1516 rated models, from 312 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| QIMMA - 3LM STEM | 95.33 | Accuracy (%, log-likelihood multiple choice) | 100 |
| Yandex AliceAI Cards - Codeforces C++ (pass@8) | 73.7 | Pass@8 (%) | 100 |
| Yandex AliceAI Cards - EduBench Math | 80 | Score (%) | 100 |
| Yandex AliceAI Cards - GPQA | 47.1 | Accuracy (%) | 100 |
| Yandex AliceAI Cards - GSM8K (CoT) | 90.4 | Accuracy (%) | 100 |
| Yandex AliceAI Cards - HumanEval | 88.3 | Score (%) | 100 |
| Yandex AliceAI Cards - MBPP | 75.4 | Score (%) | 100 |
| Yandex AliceAI Cards - MMLU | 84.4 | Accuracy (%) | 100 |
| Yandex AliceAI Cards - RULER 128K | 90.1 | Score (%) | 100 |
| Yandex AliceAI Cards - YExtract | 42.4 | Score (%) | 100 |
| EuroEval French NLU | 77.13 | NLU Average Score (%) | 98.9 |
| EuroEval Italian NLU - ScaLA IT | 67.65 | Linguistic acceptability Score (%) | 98.4 |
Interactive version: theaggregate.ai/model?slug=qwen-3-5-35b-a3b-base · How It Works · Data refreshed daily, snapshot 2026-09-24.