Qwen3-30B-A3B-Instruct Jul 2025
Qwen3-30B-A3B-Instruct.
30
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Qwen Jul 2025 Qwen3-30B-A3B-Instruct
Jul 2025 snapshot Open weights
135.2
#142 of 256
Epoch Capabilities Index
- GPQA Diamond 55.6% #147 /229
- ProfBench 40% #64 /72
- StructEval 71.6% #20 /28
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ALRAGE | 80.4% | #5 of 39 | Official leaderboard | |
| Encyclo-K | 36.7% | #8 of 23 | Official leaderboard | |
| LMArena (Elo) | 1,382 | #153 of 335 | Official leaderboard | |
| ArbMMLU-HT | 37.4% | #23 of 39 | Official leaderboard | |
| Open Arabic LLM Leaderboard v2 | 48.6% | #24 of 39 | Official leaderboard | |
| AlGhafa | 50.9% | #25 of 39 | Official leaderboard | |
| ArabicMMLU | 35.9% | #29 of 39 | Official leaderboard | |
| MadinahQA | 25.1% | #36 of 39 | Official leaderboard | |
| Judgemark v2.1 | 34.6 | #41 of 42 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SciArena | 1,010 | #17 of 34 | Official leaderboard | |
| OTIS Mock AIME 2024-2025 | 62.2% | #123 of 212 | Epoch AI run | |
| GPQA Diamond | 55.6% | #147 of 229 | Epoch AI run | |
| DTBench | 67.2% | #123 of 180 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ArtifactsBench (Gemini 2.5 Pro 0506 judge) | 40.1% | #8 of 18 | Official leaderboard | |
| BigCodeArena | 1,005 | #7 of 13 | Official leaderboard | |
| StructEval | 71.6% | #20 of 28 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BFCL V4 | 41.4% | #26 of 65 | Official leaderboard | |
| MCP-Bench | 0.63 | #7 of 15 | Official leaderboard | |
| OdysseyArena Repo System | 26.7% | #4 of 8 | Official leaderboard | |
| OdysseyArena Energy Dispatch | 0% | #5 of 8 | Official leaderboard | |
| OdysseyArena AI Trading | 4.76% | #6 of 8 | Official leaderboard | |
| OdysseyArena Turn on Lights | 11.7% | #6 of 8 | Official leaderboard | |
| LisanBench | 74.7 | #118 of 133 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ProfBench | 40% | #64 of 72 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AraTrust | 72.7% | #21 of 39 | Official leaderboard | |
| BrowseSafe++ | 65.7% | #13 of 15 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LMCA | 22.4% | #106 of 147 | Official leaderboard | |
| Conceptual Reasoning Index | 31.6 | #113 of 151 | Official leaderboard | |
| ACCoRD | 45.4 | #121 of 152 | Official leaderboard | |
| Political Science LLM Benchmark | 63% | #38 of 42 | Official leaderboard |
Qwen's models over time
Qwen's models in color, against every other model on the Epoch Capabilities Index, by release date.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)