Qwen 3 30B A3B 2507 Instruct: benchmark results
Alibaba's July 2025 refresh of its open 30B A3B MoE (3.3.3B active), non-thinking instruct variant with 256K native context. Provider: Alibaba. Released 2025-07-30. Access: Open.
Aggregate Rating 1586 ± 6, rank #488 of 1656 rated models, from 1121 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CoopEval - Reputation (First-Order) | 0.4 | Normalized mean payoff (open scale; below 0 when exploited, | 100 |
| CoopEval - Reputation (Higher-Order) | 0.32 | Normalized mean payoff (open scale; below 0 when exploited, | 100 |
| GPSBench (Applied) - Boundary Analysis | 99.7 | Score (%) on the boundary analysis (group four to six named | 100 |
| LIBRA - LibrusecHistory | 96.12 | Dataset Total Score (%) | 100 |
| LIBRA - LibrusecMHQA * | 51 | Dataset Total Score (%) | 100 |
| LIBRA - MatreshkaNames * | 81.2 | Dataset Total Score (%) | 100 |
| LIBRA - ru2WikiMultihopQA * | 66.63 | Dataset Total Score (%) | 100 |
| LIBRA - ruBABILongQA1 | 80.5 | Dataset Total Score (%) | 100 |
| LIBRA - ruBABILongQA2 | 64.72 | Dataset Total Score (%) | 100 |
| LIBRA - ruBABILongQA3 * | 38.38 | Dataset Total Score (%) | 100 |
| LIBRA - ruBABILongQA4 | 70.98 | Dataset Total Score (%) | 100 |
| LIBRA - ruBABILongQA5 | 82.95 | Dataset Total Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=qwen-3-30b-a3b-2507-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.