Qwen-3-Swallow-30B-A3B-SFT-v0.2: benchmark results
Provider: Alibaba.
Aggregate Rating 1577 ± 34, rank #1296 of 10916 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Swallow - Japanese MT-Bench - Roleplay | 70.8 | Judge Score (normalized, %) | 80.6 |
| Swallow - Japanese MT-Bench - Coding | 77.1 | Judge Score (normalized, %) | 61.2 |
| Swallow - English MT-Bench - Math | 97.4 | Judge Score (normalized, %) | 59.7 |
| Swallow - English MT-Bench - Reasoning | 82.5 | Judge Score (normalized, %) | 56.7 |
| Swallow - English MT-Bench - Roleplay | 73.3 | Judge Score (normalized, %) | 56.7 |
| Swallow - Post-trained English - LiveCodeBench | 50.5 | Pass@1 (%) | 55.2 |
| Swallow - Japanese MT-Bench - Humanities | 63.2 | Judge Score (normalized, %) | 53.7 |
| Swallow - Japanese MT-Bench - STEM | 64.8 | Judge Score (normalized, %) | 53.7 |
| Swallow - Japanese MT-Bench - Average | 71.4 | Judge Score (normalized, %) | 53 |
| Swallow - Post-trained Japanese - WMT20 En-Ja | 24 | BLEU | 53 |
| Swallow - Post-trained English - IFBench | 40.4 | Instruction-Level Strict Accuracy (%) | 52.2 |
| Swallow - Post-trained English - Average | 66.8 | Average Score (%) | 51.5 |
Interactive version: theaggregate.ai/model?slug=qwen-3-swallow-30b-a3b-sft-v0-2 · How It Works · Data refreshed daily, snapshot 2026-11-08.