Qwen-3-Swallow-30B-A3B-RL-v0.2: benchmark results
Provider: Alibaba.
Aggregate Rating 1598 ± 37, rank #1030 of 10916 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Swallow - Post-trained Japanese - PolyMath High and Top | 52.4 | Accuracy (%) | 80.6 |
| Swallow - Japanese MT-Bench - Reasoning | 79.4 | Judge Score (normalized, %) | 79.9 |
| Swallow - Post-trained English - AIME | 81.7 | Accuracy (%) | 77.6 |
| Swallow - Post-trained English - MATH-500 | 97.6 | Accuracy (%) | 77.6 |
| Swallow - Japanese MT-Bench - Math | 97.8 | Judge Score (normalized, %) | 76.9 |
| Swallow - Post-trained English - LiveCodeBench | 61.8 | Pass@1 (%) | 76.1 |
| Swallow - English MT-Bench - Reasoning | 85.4 | Judge Score (normalized, %) | 73.1 |
| Swallow - Japanese MT-Bench - STEM | 69 | Judge Score (normalized, %) | 73.1 |
| Swallow - Japanese MT-Bench - Average | 74.1 | Judge Score (normalized, %) | 70.1 |
| Swallow - Post-trained Japanese - JHumanEval | 91.7 | Pass@1 (%) | 70.1 |
| Swallow - English MT-Bench - Math | 98.5 | Judge Score (normalized, %) | 69.4 |
| Swallow - Post-trained Japanese - Average | 54.6 | Average Score (%) | 68.7 |
Interactive version: theaggregate.ai/model?slug=qwen-3-swallow-30b-a3b-rl-v0-2 · How It Works · Data refreshed daily, snapshot 2026-11-08.