Model comparison
GPT-4.5 vs Qwen3-4B
GPT-4.5 is the stronger model overall, scoring 36.9 to 31.5 on the Noometry Index.
Last verified . 2 shared benchmarks.
Alibaba (Qwen)Summary
- They share 2 benchmarks with published results for both. GPT-4.5 scores higher in 2 categories and Qwen3-4B in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3-4B leads 18.5 to 14.1.
- The biggest single-benchmark swing is GPQA Diamond: 68.7% for GPT-4.5 and 52.3% for Qwen3-4B.
- Qwen3-4B has downloadable open weights; the other is API-only.
Side by side
| GPT-4.5 | Qwen3-4B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 36.9 | 31.5 |
| Released | 2025-02-27 | 2025-04-29 |
| Weights | Proprietary | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 42 | 6 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
GPT-4.5: 41.8 (#102), Qwen3-4B: —
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| Aider Polyglot | 44.9% | — |
| WeirdML | 39.4% | — |
| LiveBench Coding | 75.2% | — |
| LMArena Coding | 1397 | — |
Agentic & Tool Use Too close to call
GPT-4.5: 27.6 (#92), Qwen3-4B: 27.3 (#94)
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 35.7% |
| Cybench | 17.5% | — |
Reasoning Qwen3-4B leads
GPT-4.5: 14.1 (#320), Qwen3-4B: 18.5 (#264)
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| ARC-AGI-2 | 0.8% | — |
| SimpleBench | 34.5% | — |
| ARC-AGI-1 | 10.3% | — |
| Chess Puzzles | — | 4% |
| EnigmaEval | 3.2% | — |
| LiveBench Reasoning | 71.1% | — |
| LMArena Hard Prompts | 1404 | — |
| LiveBench Data Analysis | 64.3% | — |
| Epoch Capabilities Index | 136.74 | — |
| ForecastBench | 61.7 | — |
| LiveBench | 69% | — |
Math GPT-4.5 leads
GPT-4.5: 31.0 (#200), Qwen3-4B: 28.5 (#228)
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 52.2% |
| MathArena Final-Answer Competitions | — | 38.5% |
| LiveBench Math | 69.3% | — |
| LMArena Math | 1412 | — |
| MATH Level 5 | 78.6% | — |
Knowledge Too close to call
GPT-4.5: 32.4 (#205), Qwen3-4B: 33.0 (#198)
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| GPQA Diamond | 68.7% | 52.3% |
| Humanity's Last Exam | 5.4% | — |
| Confabulations | 13.6% | — |
| Vectara Hallucination Rate | — | 5.7% |
| LMArena Expert | 1394 | — |
Multimodal Not comparable
GPT-4.5: 37.0 (#68), Qwen3-4B: —
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| LMArena Vision | 1195 | — |
| VPCT | 45% | — |
Multilingual Not comparable
GPT-4.5: 52.7 (#78), Qwen3-4B: —
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| LMArena Non-English | 1413 | — |
| LMArena Chinese | 1421 | — |
| LMArena French | 1417 | — |
| LMArena German | 1457 | — |
| LMArena Japanese | 1416 | — |
| LMArena Korean | 1392 | — |
| LMArena Russian | 1419 | — |
Instruction Following Not comparable
GPT-4.5: 72.8 (#129), Qwen3-4B: —
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| LiveBench Instruction Following | 72.3% | — |
| LMArena Instruction Following | 1404 | — |
Long Context Not comparable
GPT-4.5: 40.4 (#150), Qwen3-4B: —
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| Fiction.LiveBench | 63.9% | — |
| LMArena Longer Query | 1406 | — |
Writing & Preference Not comparable
GPT-4.5: 57.1 (#129), Qwen3-4B: —
| Benchmark | GPT-4.5 | Qwen3-4B |
|---|---|---|
| LMArena Text | 1417 | — |
| LMArena Creative Writing | 1394 | — |
| Short-Story Creative Writing | 75.6% | — |
| EQ-Bench Creative Writing | 1258 | — |
| LMArena Multi-Turn | 1444 | — |
| LiveBench Language | 61.5% | — |
Frequently asked questions
Is GPT-4.5 better than Qwen3-4B?
GPT-4.5 is the stronger model overall, scoring 36.9 to 31.5 on the Noometry Index.
How many benchmarks do GPT-4.5 and Qwen3-4B share?
2 benchmarks have published results for both models. GPT-4.5 has 42 scored results on Noometry and Qwen3-4B has 6.