Model comparison
Grok 3 vs Qwen2.5-Coder-32B
Grok 3 is the stronger model overall, scoring 39.6 to 32.9 on the Noometry Index.
Last verified . 14 shared benchmarks.
Alibaba (Qwen)Summary
- They share 14 benchmarks with published results for both. Grok 3 scores higher in 7 categories and Qwen2.5-Coder-32B in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Grok 3 leads 41.5 to 22.4.
- The biggest single-benchmark swing is Aider Polyglot: 53.3% for Grok 3 and 16.4% for Qwen2.5-Coder-32B.
- Qwen2.5-Coder-32B has downloadable open weights; the other is API-only.
Side by side
| Grok 3 | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | xAI | Alibaba (Qwen) |
| Noometry Index | 39.6 | 32.9 |
| Released | 2025-04-09 | 2024-09-18 |
| Weights | Proprietary | Open |
| Context window | — | 33K |
| Max output | — | 29K |
| Input $ / M tokens | — | $0.66 |
| Output $ / M tokens | — | $1 |
| Results tracked | 40 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 3 leads
Grok 3: 41.5 (#113), Qwen2.5-Coder-32B: 22.4 (#328)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| Aider Polyglot | 53.3% | 16.4% |
| LMArena Coding | 1432 | 1276 |
| SWE-bench Verified (bash only) | — | 9% |
| WeirdML | 37.2% | — |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Agentic & Tool Use Not comparable
Grok 3: 30.2 (#69), Qwen2.5-Coder-32B: —
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| BALROG | 29.5% | — |
Reasoning Qwen2.5-Coder-32B leads
Grok 3: 13.8 (#322), Qwen2.5-Coder-32B: 20.4 (#220)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1433 | 1251 |
| Epoch Capabilities Index | 138.34 | 119.41 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 36.1% | — |
| Kagi LLM Benchmark | 61.3% | — |
| ARC-AGI-1 | 5.5% | — |
| LiveBench Reasoning | — | 42.1% |
| LiveBench Data Analysis | — | 49.9% |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
| WinoGrande | — | 80.8% |
Math Grok 3 leads
Grok 3: 36.2 (#122), Qwen2.5-Coder-32B: 30.8 (#203)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1390 | 1251 |
| OTIS Mock AIME 2024-2025 | 55.6% | — |
| Omni-MATH | 46.4% | — |
| LiveBench Math | — | 46.6% |
| MATH Level 5 | 88.7% | — |
| FrontierMath (Feb 2025 set) | 3.8% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
| GSM8K | — | 93% |
Knowledge Grok 3 leads
Grok 3: 46.0 (#76), Qwen2.5-Coder-32B: 32.9 (#200)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1420 | 1221 |
| GPQA Diamond | 75.8% | — |
| MMLU-Pro | 78.8% | — |
| Confabulations | 14.2% | — |
| Vectara Hallucination Rate | 5.8% | — |
| GPQA (HELM) | 65% | — |
| ARC (AI2) Challenge | — | 70.5% |
| MMLU | — | 79.1% |
Multilingual Grok 3 leads
Grok 3: 52.5 (#81), Qwen2.5-Coder-32B: 38.0 (#230)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1411 | 1205 |
| LMArena Chinese | 1448 | 1222 |
| LMArena Russian | 1416 | 1228 |
| LMArena French | 1460 | — |
| LMArena German | 1431 | — |
| LMArena Japanese | 1389 | — |
| LMArena Korean | 1374 | — |
| LMArena Spanish | 1417 | — |
Instruction Following Grok 3 leads
Grok 3: 75.2 (#69), Qwen2.5-Coder-32B: 61.6 (#240)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1409 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
| IFEval | 88.4% | — |
Long Context Too close to call
Grok 3: 38.7 (#187), Qwen2.5-Coder-32B: 37.9 (#204)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1439 | 1251 |
| Fiction.LiveBench | 58.3% | — |
Writing & Preference Grok 3 leads
Grok 3: 56.0 (#136), Qwen2.5-Coder-32B: 41.8 (#235)
| Benchmark | Grok 3 | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1426 | 1230 |
| LMArena Creative Writing | 1414 | 1174 |
| LMArena Multi-Turn | 1424 | 1222 |
| Short-Story Creative Writing | 76.4% | — |
| EQ-Bench Creative Writing | 1186 | — |
| WildBench | 84.9% | — |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is Grok 3 better than Qwen2.5-Coder-32B?
Grok 3 is the stronger model overall, scoring 39.6 to 32.9 on the Noometry Index.
Is Grok 3 or Qwen2.5-Coder-32B better for coding?
Grok 3 scores higher on coding benchmarks: 41.5 versus 22.4 in the Noometry coding category.
How many benchmarks do Grok 3 and Qwen2.5-Coder-32B share?
14 benchmarks have published results for both models. Grok 3 has 40 scored results on Noometry and Qwen2.5-Coder-32B has 31.