Model comparison
GPT-4o vs Qwen2.5-Coder-32B
Qwen2.5-Coder-32B is the stronger model overall, scoring 33.4 to 28.6 on the Noometry Index.
Last verified . 27 shared benchmarks.
Alibaba (Qwen)Summary
- They share 27 benchmarks with published results for both. GPT-4o scores higher in 5 categories and Qwen2.5-Coder-32B in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen2.5-Coder-32B leads 33.3 to 10.6.
- The biggest single-benchmark swing is Aider Polyglot: 45.3% for GPT-4o and 16.4% for Qwen2.5-Coder-32B.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $2.50 / $10 for GPT-4o.
- GPT-4o accepts more context: 128K tokens versus 33K.
- Qwen2.5-Coder-32B has downloadable open weights; the other is API-only.
Side by side
| GPT-4o | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 28.6 | 33.4 |
| Released | 2024-05-13 | 2024-09-18 |
| Weights | Proprietary | Open |
| Context window | 128K | 33K |
| Max output | 16K | 29K |
| Input $ / M tokens | $2.50 | $0.66 |
| Output $ / M tokens | $10 | $1 |
| Results tracked | 72 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4o leads
GPT-4o: 24.8 (#328), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| SWE-bench Verified (bash only) | 21.6% | 9% |
| Aider Polyglot | 45.3% | 16.4% |
| BigCodeBench Instruct | 51.1% | 49% |
| LiveBench Coding | 51.4% | 56.9% |
| LMArena Coding | 1297 | 1276 |
| BigCodeBench Complete | 61.1% | 58% |
| HumanEval+ | 87.2% | 87.2% |
| MBPP+ | 72.2% | 77% |
| SWE-bench Verified | 31% | — |
| GSO | 0% | — |
| WeirdML | 25.1% | — |
| CadEval | 26% | — |
Agentic & Tool Use Not comparable
GPT-4o: 21.0 (#141), Qwen2.5-Coder-32B: —
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| GDPval | 9.9% | — |
| TheAgentCompany | 8.6% | — |
| Cybench | 12.5% | — |
| BALROG | 32.3% | — |
| LMArena Search | 1006 | — |
| METR Time Horizons | 40.8% | — |
Reasoning Qwen2.5-Coder-32B leads
GPT-4o: 9.4 (#343), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LiveBench Reasoning | 55.8% | 42.1% |
| LMArena Hard Prompts | 1281 | 1251 |
| LiveBench Data Analysis | 60.9% | 49.9% |
| Epoch Capabilities Index | 128.97 | 119.49 |
| LiveBench | 55.3% | 46.2% |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 17.8% | — |
| ARC-AGI-1 | 4.5% | — |
| CritPt | 0% | — |
| Chess Puzzles | 13% | — |
| EnigmaEval | 0.8% | — |
| DTBench | 64.5% | — |
| LMCA | 16.6% | — |
| ForecastBench | 57.7 | — |
| HellaSwag | — | 83% |
| WinoGrande | — | 80.8% |
Math Qwen2.5-Coder-32B leads
GPT-4o: 10.6 (#312), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LiveBench Math | 49.5% | 46.6% |
| LMArena Math | 1285 | 1251 |
| FrontierMath (Tiers 1-3) | 0.4% | — |
| OTIS Mock AIME 2024-2025 | 6.4% | — |
| Omni-MATH | 29.3% | — |
| MATH Level 5 | 53.3% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
| GSM8K | — | 93% |
Knowledge Qwen2.5-Coder-32B leads
GPT-4o: 28.8 (#242), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1250 | 1221 |
| MMLU | 88.1% | 79.1% |
| GPQA Diamond | 49.2% | — |
| Humanity's Last Exam | 2.7% | — |
| SimpleQA Verified | 26% | — |
| MMLU-Pro | 71.3% | — |
| Confabulations | 15.3% | — |
| Vectara Hallucination Rate | 9.6% | — |
| GPQA (HELM) | 52% | — |
| ARC (AI2) Challenge | — | 70.5% |
Multimodal Not comparable
GPT-4o: 34.5 (#91), Qwen2.5-Coder-32B: —
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Vision | 1137 | — |
| Video-MME | 71.9% | — |
| GeoBench | 71% | — |
| VPCT | 40% | — |
| ScienceQA | 88.5% | — |
Multilingual GPT-4o leads
GPT-4o: 43.2 (#186), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1283 | 1205 |
| LMArena Chinese | 1277 | 1222 |
| LMArena Russian | 1286 | 1228 |
| LMArena French | 1304 | — |
| LMArena German | 1282 | — |
| LMArena Japanese | 1257 | — |
| LMArena Korean | 1234 | — |
| LMArena Spanish | 1292 | — |
Instruction Following GPT-4o leads
GPT-4o: 66.6 (#207), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LiveBench Instruction Following | 68.6% | 58.7% |
| LMArena Instruction Following | 1278 | 1223 |
| IFEval | 81.7% | — |
Long Context GPT-4o leads
GPT-4o: 39.4 (#179), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1289 | 1251 |
| Fiction.LiveBench | 66.7% | — |
Writing & Preference GPT-4o leads
GPT-4o: 52.6 (#166), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | GPT-4o | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1300 | 1230 |
| LMArena Creative Writing | 1292 | 1174 |
| LMArena Multi-Turn | 1302 | 1222 |
| LiveBench Language | 47.6% | 23.3% |
| Short-Story Creative Writing | 81.8% | — |
| WildBench | 82.8% | — |
Frequently asked questions
Is GPT-4o better than Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is the stronger model overall, scoring 33.4 to 28.6 on the Noometry Index.
Which is cheaper, GPT-4o or Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; GPT-4o lists at $2.50 and $10.
Is GPT-4o or Qwen2.5-Coder-32B better for coding?
GPT-4o scores higher on coding benchmarks: 24.8 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
GPT-4o does, with 128K tokens against 33K.
How many benchmarks do GPT-4o and Qwen2.5-Coder-32B share?
27 benchmarks have published results for both models. GPT-4o has 72 scored results on Noometry and Qwen2.5-Coder-32B has 31.