Model comparison
GPT-5.4 mini vs Qwen2.5 72B Instruct
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 31.9 on the Noometry Index.
Last verified . 24 shared benchmarks.
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. GPT-5.4 mini 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-5.4 mini leads 45.5 to 19.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for GPT-5.4 mini and 8.1% for Qwen2.5 72B Instruct.
- GPT-5.4 mini is cheaper at $0.75 / $4.50 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- GPT-5.4 mini accepts more context: 400K tokens versus 131K.
- Qwen2.5 72B Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 mini | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 45.0 | 31.9 |
| Released | 2026-03-17 | 2024-09 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 8K |
| Input $ / M tokens | $0.75 | $1.40 |
| Output $ / M tokens | $4.50 | $5.60 |
| Results tracked | 46 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 mini leads
GPT-5.4 mini: 45.2 (#72), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| WeirdML | 60.3% | 16% |
| LMArena Coding | 1438 | 1292 |
| FrontierCode | 27% | — |
| LMArena WebDev | 1397 | — |
| SciCode | 49.9% | — |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use GPT-5.4 mini leads
GPT-5.4 mini: 29.9 (#81), Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| DeepResearch Bench | 36.3% | — |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning GPT-5.4 mini leads
GPT-5.4 mini: 30.4 (#85), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1424 | 1271 |
| DTBench | 80% | 62.9% |
| LMCA | 40.8% | 13.4% |
| Epoch Capabilities Index | 148.84 | 129 |
| ForecastBench | 57 | 57.5 |
| ARC-AGI-2 | 18.9% | — |
| Kagi LLM Benchmark | 37.9% | — |
| NYT Connections (extended) | 61.8% | — |
| ARC-AGI-1 | 63.7% | — |
| CritPt | 10% | — |
| Chess Puzzles | 24% | — |
| Thematic Generalization | 61.7% | — |
| Mystery Game Puzzles | 11% | — |
| BIG-Bench Hard | — | 79.8% |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math GPT-5.4 mini leads
GPT-5.4 mini: 45.5 (#75), Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 8.1% |
| LMArena Math | 1419 | 1283 |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 9.8% | — |
| ProofBench | 21% | — |
| Omni-MATH | — | 33% |
| MATH Level 5 | — | 63.2% |
| FrontierMath (Feb 2025 set) | 28.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5.4 mini leads
GPT-5.4 mini: 51.5 (#67), Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 86.9% | 49.1% |
| LMArena Expert | 1435 | 1245 |
| SimpleQA Verified | 29.4% | — |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| Vectara Hallucination Rate | 5.5% | — |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multimodal Not comparable
GPT-5.4 mini: 39.7 (#56), Qwen2.5 72B Instruct: —
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Vision | 1245 | — |
Multilingual GPT-5.4 mini leads
GPT-5.4 mini: 51.9 (#96), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1405 | 1252 |
| LMArena Chinese | 1446 | 1272 |
| LMArena French | 1440 | 1280 |
| LMArena German | 1409 | 1234 |
| LMArena Japanese | 1374 | 1180 |
| LMArena Korean | 1368 | 1188 |
| LMArena Russian | 1417 | 1264 |
| LMArena Spanish | 1405 | 1256 |
Instruction Following GPT-5.4 mini leads
GPT-5.4 mini: 74.1 (#102), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1405 | 1254 |
| IFEval | — | 80.6% |
Long Context GPT-5.4 mini leads
GPT-5.4 mini: 43.0 (#112), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1407 | 1282 |
Writing & Preference GPT-5.4 mini leads
GPT-5.4 mini: 64.0 (#58), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | GPT-5.4 mini | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1412 | 1269 |
| LMArena Creative Writing | 1370 | 1221 |
| LMArena Multi-Turn | 1429 | 1272 |
| EQ-Bench Creative Writing | 1665 | — |
| WildBench | — | 80.2% |
Frequently asked questions
Is GPT-5.4 mini better than Qwen2.5 72B Instruct?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 31.9 on the Noometry Index.
Which is cheaper, GPT-5.4 mini or Qwen2.5 72B Instruct?
GPT-5.4 mini is cheaper. It lists at $0.75 per million input tokens and $4.50 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is GPT-5.4 mini or Qwen2.5 72B Instruct better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 mini does, with 400K tokens against 131K.
How many benchmarks do GPT-5.4 mini and Qwen2.5 72B Instruct share?
24 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Qwen2.5 72B Instruct has 43.