Qwen3-235B-A22B-awq: benchmark results
Provider: Alibaba.
Aggregate Rating 1587 ± 33, rank #1152 of 10916 rated models, from 28 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| TurkBench - STS | 93.34 | Pearson correlation x100 | 95.9 |
| TurkBench - Topic Detection | 86.41 | Accuracy (%) | 86.1 |
| TurkBench - Metaphors and Idioms | 79.33 | Accuracy (%) | 81.1 |
| TurkBench - Toxicity | 98.7 | LLM-as-a-judge score (0-100) | 75.7 |
| TurkBench - Instruction Following | 91.19 | LLM-as-a-judge score (0-100) | 67.6 |
| TurkBench - MMLU Topics | 61.3 | Accuracy (%) | 64.9 |
| TurkBench - Parts of Speech | 77.73 | Accuracy (%) | 64.9 |
| TurkBench - Reading Comprehension | 93.96 | LLM-as-a-judge score (0-100) | 64.9 |
| TurkBench - Math | 33.8 | Accuracy (%) | 56.8 |
| AgentEHR (ReAct) - Procedures | 27.39 | F1 (%; predicted procedures label set against the next recor | 50 |
| TurkBench - Named Entity Recognition | 55.83 | Accuracy (%) | 48.6 |
| TurkBench - Faithfulness | 86.08 | LLM-as-a-judge score (0-100) | 45.9 |
Interactive version: theaggregate.ai/model?slug=qwen3-235b-a22b-awq · How It Works · Data refreshed daily, snapshot 2026-11-08.