QwQ-32B
QwQ-32B.
55
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Qwen Mar 2025 QwQ-32B
Open weights Open model
136.5
#135 of 256
Epoch Capabilities Index
- GPQA Diamond 65.3% #130 /229
- LiveSQLBench 18.2% #29 /36
- Encyclo-K 41% #6 /23
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 | |
|---|---|---|---|---|
| Fiction.LiveBench | 83.3% | #10 of 57 | Fiction.live leaderboard, via Epoch AI | |
| Encyclo-K | 41% | #6 of 23 | Official leaderboard | |
| Stick to Your Role! | 0.77 | #5 of 15 | Official leaderboard | |
| AgentIF | 58.1% | #4 of 12 | Official leaderboard | |
| LLM Confabulation Leaderboard | 15.6% | #23 of 58 | Official leaderboard | |
| MuSR | 42.1% | #41 of 98 | Official leaderboard | |
| MMLU-Pro | 69.1% | #89 of 212 | Official leaderboard | |
| WritingBench | 72.4 | #22 of 45 | Official leaderboard | |
| MathIF | 40.2% | #6 of 12 | Official leaderboard | |
| LongBench v2 | 48.9% | #17 of 33 | Official leaderboard | |
| EQ-Bench 3 | 1,185 | #44 of 78 | Official leaderboard | |
| Wer wird Millionär? LLM Benchmark | 2178.0 | #44 of 77 | Official leaderboard | |
| LMArena (Elo) | 1,336 | #194 of 335 | Official leaderboard | |
| Arena-Hard v2.0 | 43.5% | #14 of 23 | Official leaderboard | |
| FaithJudge | 24.7% | #24 of 37 | Official leaderboard | |
| ALRAGE | 60.7% | #28 of 39 | Official leaderboard | |
| Creative Writing v3 | 56.5 | #94 of 120 | Official leaderboard | |
| TableBench | 43.9% | #8 of 10 | Official leaderboard | |
| MadinahQA | 29.4% | #32 of 39 | Official leaderboard | |
| Longform Creative Writing (EQ-Bench) | 36.9 | #102 of 119 | Official leaderboard | |
| Open Arabic LLM Leaderboard v2 | 32.8% | #37 of 39 | Official leaderboard | |
| ArabicMMLU | 29.8% | #38 of 39 | Official leaderboard | |
| ArbMMLU-HT | 23.4% | #38 of 39 | Official leaderboard | |
| AlGhafa | 31.3% | #39 of 39 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| PolyMath | 45.9% | #7 of 38 | Official leaderboard | |
| Misguided Attention | 56.1% | #11 of 45 | Official leaderboard | |
| GPQA Diamond | 65.3% | #130 of 229 | Epoch AI run | |
| OTIS Mock AIME 2024-2025 | 59.2% | #125 of 212 | Epoch AI run | |
| SciArena | 978 | #23 of 34 | Official leaderboard | |
| HMMT 2025 | 47.5% | #40 of 53 | Official leaderboard | |
| AIME 2025 | 65.8% | #41 of 53 | Official leaderboard | |
| PHYBench | 15.8% | #22 of 28 | Official leaderboard | |
| IneqMath | 2.00% | #27 of 34 | Official leaderboard | |
| USAMO 2025 | 2.98% | #8 of 10 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BigO(Bench) | 40.4% | #4 of 18 | Official leaderboard | |
| BigCodeBench-Hard | 29% | #23 of 96 | Official leaderboard | |
| ArtifactsBench (Gemini 2.5 Pro 0506 judge) | 40.8% | #7 of 18 | Official leaderboard | |
| OIBench | 10.7% | #6 of 9 | Official leaderboard | |
| Aider Polyglot | 20.9% | #35 of 46 | Official leaderboard | |
| LiveSQLBench | 18.2% | #29 of 36 | Official leaderboard | |
| Spider 2.0 | 8.96% | #22 of 23 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Step Race | 3 | #17 of 59 | Official leaderboard | |
| Berkeley Function Calling Leaderboard (BFCL) | 58.3% | #28 of 67 | Official leaderboard | |
| BrowseComp-ZH | 11.1% | #6 of 9 | Official leaderboard | |
| GAIA | 22.3% | #53 of 64 | Official leaderboard | |
| Elimination Game | 2 | #53 of 56 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BRIDGE | 48.2 | #14 of 39 | Official leaderboard | |
| LEXam | 44.4% | #15 of 30 | Official leaderboard | |
| BizFinBench | 65.8% | #7 of 13 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| VRBench | 52.5% | #11 of 19 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BaxBench | 25.5% | #19 of 34 | Official leaderboard | |
| Lakera B3 (Backbone Breaker Benchmark) | 52.2% | #22 of 35 | Official leaderboard | |
| AIRTBench | 0.57% | #8 of 9 | Official leaderboard | |
| AraTrust | 31.2% | #39 of 39 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Open LLM Leaderboard v2 | 12.2 | #83 of 99 | 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)
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.