Model comparison
GPT-5.4 mini vs Qwen Turbo
GPT-5.4 mini is the stronger model overall, scoring 44.8 to 26.8 on the Noometry Index. Qwen Turbo costs 19× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Alibaba (Qwen)Summary
- They share 2 benchmarks with published results for both. GPT-5.4 mini scores higher in 2 categories and Qwen Turbo in 0 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 mini leads 45.4 to 14.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for GPT-5.4 mini and 6.1% for Qwen Turbo.
- Qwen Turbo is cheaper at $0.05 / $0.20 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
- Qwen Turbo accepts more context: 1M tokens versus 400K.
Side by side
| GPT-5.4 mini | Qwen Turbo | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 44.8 | 26.8 |
| Released | 2026-03-17 | 2024-11-01 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1M |
| Max output | 128K | 16K |
| Input $ / M tokens | $0.75 | $0.05 |
| Output $ / M tokens | $4.50 | $0.20 |
| Results tracked | 46 | 3 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
GPT-5.4 mini: 45.0 (#67), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| FrontierCode | 27% | — |
| LMArena WebDev | 1397 | — |
| SciCode | 49.9% | — |
| WeirdML | 60.3% | — |
| LMArena Coding | 1438 | — |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use Not comparable
GPT-5.4 mini: 29.6 (#74), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| DeepResearch Bench | 36.3% | — |
Reasoning Not comparable
GPT-5.4 mini: 30.3 (#77), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| 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% | — |
| LMArena Hard Prompts | 1424 | — |
| Mystery Game Puzzles | 11% | — |
| DTBench | 80% | — |
| LMCA | 40.8% | — |
| Epoch Capabilities Index | 148.83 | — |
| ForecastBench | 57 | — |
Math GPT-5.4 mini leads
GPT-5.4 mini: 45.4 (#63), Qwen Turbo: 14.5 (#291)
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 6.1% |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 9.8% | — |
| ProofBench | 21% | — |
| LMArena Math | 1418 | — |
| MATH Level 5 | — | 56.2% |
| FrontierMath (Feb 2025 set) | 28.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5.4 mini leads
GPT-5.4 mini: 51.4 (#62), Qwen Turbo: 22.2 (#267)
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| GPQA Diamond | 86.9% | 41.8% |
| SimpleQA Verified | 29.4% | — |
| Vectara Hallucination Rate | 5.5% | — |
| LMArena Expert | 1434 | — |
Multimodal Not comparable
GPT-5.4 mini: 39.0 (#51), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| LMArena Vision | 1245 | — |
Multilingual Not comparable
GPT-5.4 mini: 52.1 (#94), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| LMArena Non-English | 1405 | — |
| LMArena Chinese | 1446 | — |
| LMArena French | 1440 | — |
| LMArena German | 1407 | — |
| LMArena Japanese | 1375 | — |
| LMArena Korean | 1369 | — |
| LMArena Russian | 1417 | — |
| LMArena Spanish | 1403 | — |
Instruction Following Not comparable
GPT-5.4 mini: 74.3 (#99), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| LMArena Instruction Following | 1405 | — |
Long Context Not comparable
GPT-5.4 mini: 42.9 (#107), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| LMArena Longer Query | 1407 | — |
Writing & Preference Not comparable
GPT-5.4 mini: 64.2 (#56), Qwen Turbo: —
| Benchmark | GPT-5.4 mini | Qwen Turbo |
|---|---|---|
| LMArena Text | 1411 | — |
| LMArena Creative Writing | 1370 | — |
| EQ-Bench Creative Writing | 1665 | — |
| LMArena Multi-Turn | 1429 | — |
Frequently asked questions
Is GPT-5.4 mini better than Qwen Turbo?
GPT-5.4 mini is the stronger model overall, scoring 44.8 to 26.8 on the Noometry Index. Qwen Turbo costs 19× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 mini or Qwen Turbo?
Qwen Turbo is cheaper. It lists at $0.05 per million input tokens and $0.20 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.
Which has the bigger context window?
Qwen Turbo does, with 1M tokens against 400K.
How many benchmarks do GPT-5.4 mini and Qwen Turbo share?
2 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Qwen Turbo has 3.