# GPT-4.5 vs Qwen2-72B

> GPT-4.5 is the stronger model overall, scoring 36.9 to 29.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-5-vs-qwen2-72b
- Last updated: 2026-10-10
- Shared benchmarks: 20

## Summary

- They share 20 benchmarks with published results for both. GPT-4.5 scores higher in 8 categories and Qwen2-72B in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where GPT-4.5 leads 52.7 to 36.0.
- The biggest single-benchmark swing is MATH Level 5: 78.6% for GPT-4.5 and 39.1% for Qwen2-72B.
- Qwen2-72B has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-4.5 | Qwen2-72B |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 36.9 | 29.5 |
| Rank | 200 | 294 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Proprietary | Open |

## Coding

- GPT-4.5: 41.8 (#102)
- Qwen2-72B: 28.7 (#303)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| WeirdML | 39.4% | 11.3% |
| LMArena Coding | 1397 | 1196 |
| Aider Polyglot | 44.9% | — |
| BigCodeBench Instruct | — | 38.5% |
| LiveBench Coding | 75.2% | — |
| BigCodeBench Complete | — | 54% |

## Agentic & Tool Use

- GPT-4.5: 27.6 (#92)
- Qwen2-72B: 16.9 (#138)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| TheAgentCompany | — | 1.1% |
| Cybench | 17.5% | — |
| METR Time Horizons | — | 29.9% |

## Reasoning

- GPT-4.5: 14.1 (#320)
- Qwen2-72B: 22.2 (#184)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1404 | 1191 |
| Epoch Capabilities Index | 136.74 | 125.27 |
| ARC-AGI-2 | 0.8% | — |
| SimpleBench | 34.5% | — |
| ARC-AGI-1 | 10.3% | — |
| EnigmaEval | 3.2% | — |
| LiveBench Reasoning | 71.1% | — |
| LiveBench Data Analysis | 64.3% | — |
| ForecastBench | 61.7 | — |
| LiveBench | 69% | — |

## Math

- GPT-4.5: 31.0 (#200)
- Qwen2-72B: 27.9 (#232)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Math | 1412 | 1235 |
| MATH Level 5 | 78.6% | 39.1% |
| OTIS Mock AIME 2024-2025 | 37.8% | — |
| LiveBench Math | 69.3% | — |

## Knowledge

- GPT-4.5: 32.4 (#205)
- Qwen2-72B: 21.2 (#270)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| GPQA Diamond | 68.7% | 40.8% |
| LMArena Expert | 1394 | 1171 |
| Humanity's Last Exam | 5.4% | — |
| Confabulations | 13.6% | — |
| MMLU | — | 82.4% |

## Multimodal

- GPT-4.5: 37.0 (#68)
- Qwen2-72B: —

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Vision | 1195 | — |
| VPCT | 45% | — |

## Multilingual

- GPT-4.5: 52.7 (#78)
- Qwen2-72B: 36.0 (#239)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1413 | 1176 |
| LMArena Chinese | 1421 | 1240 |
| LMArena French | 1417 | 1169 |
| LMArena German | 1457 | 1151 |
| LMArena Japanese | 1416 | 1111 |
| LMArena Korean | 1392 | 1083 |
| LMArena Russian | 1419 | 1169 |
| LMArena Spanish | — | 1168 |

## Instruction Following

- GPT-4.5: 72.8 (#129)
- Qwen2-72B: 62.0 (#236)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1404 | 1181 |
| LiveBench Instruction Following | 72.3% | — |

## Long Context

- GPT-4.5: 40.4 (#150)
- Qwen2-72B: 36.1 (#230)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1406 | 1193 |
| Fiction.LiveBench | 63.9% | — |

## Writing & Preference

- GPT-4.5: 57.1 (#129)
- Qwen2-72B: 41.0 (#236)

| Benchmark | GPT-4.5 | Qwen2-72B |
|---|---|---|
| LMArena Text | 1417 | 1203 |
| LMArena Creative Writing | 1394 | 1181 |
| LMArena Multi-Turn | 1444 | 1195 |
| Short-Story Creative Writing | 75.6% | — |
| EQ-Bench Creative Writing | 1258 | — |
| LiveBench Language | 61.5% | — |

## FAQ

### Is GPT-4.5 better than Qwen2-72B?

GPT-4.5 is the stronger model overall, scoring 36.9 to 29.5 on the Noometry Index.

### Is GPT-4.5 or Qwen2-72B better for coding?

GPT-4.5 scores higher on coding benchmarks: 41.8 versus 28.7 in the Noometry coding category.

### How many benchmarks do GPT-4.5 and Qwen2-72B share?

20 benchmarks have published results for both models. GPT-4.5 has 42 scored results on Noometry and Qwen2-72B has 26.