Qwen3.6-35B-A3B-FP8: benchmark results
Provider: Alibaba. Released 2026-04-16. Access: Open.
Aggregate Rating 1658 ± 24, rank #301 of 1635 rated models, from 26 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| 3DVerifier - Grasp | 66.5 | Balanced accuracy (%) | 100 |
| 3DVerifier - Identity | 59 | Balanced accuracy (%) | 100 |
| 3DVerifier - Fine Geometry | 53.3 | Balanced accuracy (%) | 94.4 |
| Avalon-ToM-Bench - Intention Attribution | 96.67 | Accuracy (%) | 94.2 |
| BenCzechMark - Reading Comprehension | 74.36 | Duel Win Score (%) | 89.7 |
| Avalon-ToM-Bench - Average | 82.85 | Accuracy (%; macro-average of four abilities, balanced true/ | 88.5 |
| Avalon-ToM-Bench - Tacit Coordination | 88.57 | Accuracy (%) | 88.5 |
| BenCzechMark - Czech Language Understanding | 68.8 | Duel Win Score (%) | 85.9 |
| BenCzechMark - NER | 82.05 | Duel Win Score (%) | 84.6 |
| Avalon-ToM-Bench - Strategic Signaling | 80 | Accuracy (%) | 80.8 |
| BenCzechMark - Czech Math Reasoning | 64.1 | Duel Win Score (%) | 77.6 |
| BenCzechMark | 65.98 | Average Duel Win Score (%) | 76.9 |
Interactive version: theaggregate.ai/model?slug=qwen3-6-35b-a3b-fp8 · How It Works · Data refreshed daily, snapshot 2026-11-05.