Qwen 3 30B A3B Instruct: benchmark results
Alibaba's Qwen3-30B-A3B-Instruct-2507, a 30.5B-parameter (3.3B active) mixture-of-experts model tuned for non-thinking-mode instruction following, reasoning, coding and tool use, with 256K native context. Provider: Alibaba. Released 2025-04-29. Access: Open.
Aggregate Rating 1589 ± 18, rank #487 of 1635 rated models, from 55 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EComStage - Solution Decision | 94.25 | Task Score (%) | 100 |
| MetaGAI - Data Card Quality | 4.3 | Qualitative average (1-5): mean judge Likert score of data c | 100 |
| MetaGAI - Model Card Quality | 4.55 | Qualitative average (1-5): mean judge Likert score of model | 100 |
| Self-Driving Negotiator | 68 | Average success rate (%) | 100 |
| EComStage | 83.63 | Task Score (%) | 81.2 |
| TabVerse - QA (HTML Text) | 51.14 | Exact-match accuracy (%) answering table questions from HTML | 75 |
| Medmarks - LongHealth Task 2 | 88.58 | Score (%) | 69.9 |
| TabVerse - QA (LaTeX Text) | 48.43 | Exact-match accuracy (%) answering table questions from LaTe | 68.8 |
| Medmarks - PubMedQA | 76.6 | Score (%) | 68.2 |
| Medmarks - M-ARC | 46.33 | Score (%) | 64.8 |
| TabVerse - QA (Markdown Text) | 46.57 | Exact-match accuracy (%) answering table questions from Mark | 62.5 |
| EComStage - Scenario Route | 82.32 | Task Score (%) | 59.4 |
Interactive version: theaggregate.ai/model?slug=qwen-3-30b-a3b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-05.