Model comparison
GPT-5 vs Qwen2.5 72B Instruct
GPT-5 is the stronger model overall, scoring 50.9 to 31.9 on the Noometry Index.
Last verified . 33 shared benchmarks.
Alibaba (Qwen)Summary
- They share 33 benchmarks with published results for both. GPT-5 scores higher in 9 categories and Qwen2.5 72B Instruct in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 leads 55.0 to 19.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 91.4% for GPT-5 and 8.1% for Qwen2.5 72B Instruct.
- Qwen2.5 72B Instruct is cheaper at $1.40 / $5.60 per million input/output tokens, against $1.25 / $10 for GPT-5.
- GPT-5 accepts more context: 400K tokens versus 131K.
- Qwen2.5 72B Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-5 | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 50.9 | 31.9 |
| Released | 2025-08-07 | 2024-09 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 8K |
| Input $ / M tokens | $1.25 | $1.40 |
| Output $ / M tokens | $10 | $5.60 |
| Results tracked | 69 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
GPT-5: 50.3 (#47), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| WeirdML | 60.7% | 16% |
| LMArena Coding | 1436 | 1292 |
| SWE-bench Verified | 73.6% | — |
| SWE-bench Verified (bash only) | 65% | — |
| Aider Polyglot | 88% | — |
| LMArena WebDev | 1418 | — |
| SciCode | 42.9% | — |
| GSO | 6.9% | — |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
| ALE-Bench | 1,162 | — |
| AlgoTune | 1.67 | — |
Agentic & Tool Use GPT-5 leads
GPT-5: 33.1 (#56), Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| BALROG | 32.8% | 16.2% |
| METR Time Horizons | 69.6% | 35.8% |
| Terminal-Bench | 49.6% | — |
| GDPval | 34.8% | — |
| Remote Labor Index | 1.7% | — |
| TheAgentCompany | — | 5.7% |
| DeepResearch Bench | 49.6% | — |
| LMArena Search | 1133 | — |
Reasoning GPT-5 leads
GPT-5: 38.3 (#64), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1416 | 1271 |
| DTBench | 90.7% | 62.9% |
| LMCA | 40% | 13.4% |
| Epoch Capabilities Index | 150 | 129 |
| ForecastBench | 61.4 | 57.5 |
| ARC-AGI-2 | 9.9% | — |
| SimpleBench | 56.7% | — |
| Kagi LLM Benchmark | 72.7% | — |
| ARC-AGI-1 | 65.7% | — |
| CritPt | 12.6% | — |
| Chess Puzzles | 37% | — |
| EnigmaEval | 10.5% | — |
| EBR-Bench | 12.7% | — |
| Mystery Game Puzzles | 23% | — |
| BIG-Bench Hard | — | 79.8% |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math GPT-5 leads
GPT-5: 55.0 (#44), Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.4% | 8.1% |
| Omni-MATH | 64.7% | 33% |
| LMArena Math | 1407 | 1283 |
| MATH Level 5 | 98.1% | 63.2% |
| FrontierMath (Tiers 1-3) | 55.4% | — |
| FrontierMath Tier 4 | 22% | — |
| ProofBench | 18% | — |
| FrontierMath (Feb 2025 set) | 32.4% | — |
| FrontierMath Tier 4 (v1) | 12.5% | — |
Knowledge GPT-5 leads
GPT-5: 56.6 (#43), Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 86.2% | 49.1% |
| MMLU-Pro | 86.3% | 63.1% |
| Confabulations | 10.3% | 19.1% |
| GPQA (HELM) | 79.2% | 42.6% |
| LMArena Expert | 1419 | 1245 |
| Humanity's Last Exam | 25.3% | — |
| SimpleQA Verified | 50.1% | — |
| Vectara Hallucination Rate | 14.7% | — |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multimodal Not comparable
GPT-5: 46.8 (#13), Qwen2.5 72B Instruct: —
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Vision | 1232 | — |
| GeoBench | 81% | — |
| VPCT | 66% | — |
Multilingual GPT-5 leads
GPT-5: 51.4 (#110), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1397 | 1252 |
| LMArena Chinese | 1422 | 1272 |
| LMArena French | 1410 | 1280 |
| LMArena German | 1416 | 1234 |
| LMArena Japanese | 1409 | 1180 |
| LMArena Korean | 1360 | 1188 |
| LMArena Russian | 1406 | 1264 |
| LMArena Spanish | 1399 | 1256 |
Instruction Following GPT-5 leads
GPT-5: 73.8 (#113), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| IFEval | 87.5% | 80.6% |
| LMArena Instruction Following | 1388 | 1254 |
Long Context GPT-5 leads
GPT-5: 69.5 (#2), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1399 | 1282 |
| Fiction.LiveBench | 97.2% | — |
Writing & Preference GPT-5 leads
GPT-5: 63.4 (#65), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | GPT-5 | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1406 | 1269 |
| LMArena Creative Writing | 1365 | 1221 |
| WildBench | 85.7% | 80.2% |
| LMArena Multi-Turn | 1426 | 1272 |
| Short-Story Creative Writing | 86% | — |
| EQ-Bench Creative Writing | 1627 | — |
Frequently asked questions
Is GPT-5 better than Qwen2.5 72B Instruct?
GPT-5 is the stronger model overall, scoring 50.9 to 31.9 on the Noometry Index.
Which is cheaper, GPT-5 or Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is cheaper. It lists at $1.40 per million input tokens and $5.60 per million output tokens; GPT-5 lists at $1.25 and $10.
Is GPT-5 or Qwen2.5 72B Instruct better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
GPT-5 does, with 400K tokens against 131K.
How many benchmarks do GPT-5 and Qwen2.5 72B Instruct share?
33 benchmarks have published results for both models. GPT-5 has 69 scored results on Noometry and Qwen2.5 72B Instruct has 43.