Model comparison
GPT-5.4 mini vs Qwen3 235B-A22B
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 43.5 on the Noometry Index.
Last verified . 36 shared benchmarks.
Alibaba (Qwen)Summary
- They share 36 benchmarks with published results for both. GPT-5.4 mini scores higher in 5 categories and Qwen3 235B-A22B in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.4 mini leads 30.4 to 15.7.
- The biggest single-benchmark swing is ARC-AGI-1: 63.7% for GPT-5.4 mini and 11% for Qwen3 235B-A22B.
- Qwen3 235B-A22B is cheaper at $0.70 / $2.80 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
- GPT-5.4 mini accepts more context: 400K tokens versus 131K.
- Qwen3 235B-A22B has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 mini | Qwen3 235B-A22B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 45.0 | 43.5 |
| Released | 2026-03-17 | 2025-04 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 16K |
| Input $ / M tokens | $0.75 | $0.70 |
| Output $ / M tokens | $4.50 | $2.80 |
| Results tracked | 46 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GPT-5.4 mini: 45.2 (#72), Qwen3 235B-A22B: 44.3 (#75)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| SciCode | 49.9% | 42.4% |
| WeirdML | 60.3% | 41% |
| LMArena Coding | 1438 | 1445 |
| FrontierCode | 27% | — |
| Aider Polyglot | — | 59.6% |
| LMArena WebDev | 1397 | — |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use Qwen3 235B-A22B leads
GPT-5.4 mini: 29.9 (#81), Qwen3 235B-A22B: 33.9 (#51)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 52.1% |
| DeepResearch Bench | 36.3% | — |
| Vending-Bench 2 | — | -11.34 |
Reasoning GPT-5.4 mini leads
GPT-5.4 mini: 30.4 (#85), Qwen3 235B-A22B: 15.7 (#311)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| ARC-AGI-2 | 18.9% | 1.3% |
| Kagi LLM Benchmark | 37.9% | 69.4% |
| ARC-AGI-1 | 63.7% | 11% |
| CritPt | 10% | 0% |
| Chess Puzzles | 24% | 12% |
| LMArena Hard Prompts | 1424 | 1433 |
| Mystery Game Puzzles | 11% | 9% |
| DTBench | 80% | 80.3% |
| LMCA | 40.8% | 29.3% |
| Epoch Capabilities Index | 148.84 | 143.85 |
| ForecastBench | 57 | 59.7 |
| SimpleBench | — | 31% |
| NYT Connections (extended) | 61.8% | — |
| Thematic Generalization | 61.7% | — |
Math Qwen3 235B-A22B leads
GPT-5.4 mini: 45.5 (#75), Qwen3 235B-A22B: 50.4 (#57)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 86.7% |
| LMArena Math | 1419 | 1432 |
| FrontierMath (Feb 2025 set) | 28.3% | 8.5% |
| FrontierMath Tier 4 (v1) | 2.1% | 0% |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 9.8% | — |
| ProofBench | 21% | — |
| Omni-MATH | — | 71.8% |
| MATH Level 5 | — | 68.9% |
Knowledge GPT-5.4 mini leads
GPT-5.4 mini: 51.5 (#67), Qwen3 235B-A22B: 49.6 (#73)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| GPQA Diamond | 86.9% | 80.1% |
| SimpleQA Verified | 29.4% | 40.4% |
| Vectara Hallucination Rate | 5.5% | 9.3% |
| LMArena Expert | 1435 | 1463 |
| MMLU-Pro | — | 84.4% |
| Confabulations | — | 15.6% |
| GPQA (HELM) | — | 72.7% |
Multimodal Not comparable
GPT-5.4 mini: 39.7 (#56), Qwen3 235B-A22B: —
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Vision | 1245 | — |
Multilingual Too close to call
GPT-5.4 mini: 51.9 (#96), Qwen3 235B-A22B: 52.3 (#89)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1405 | 1409 |
| LMArena Chinese | 1446 | 1481 |
| LMArena French | 1440 | 1445 |
| LMArena German | 1409 | 1433 |
| LMArena Japanese | 1374 | 1399 |
| LMArena Korean | 1368 | 1391 |
| LMArena Russian | 1417 | 1411 |
| LMArena Spanish | 1405 | 1430 |
Instruction Following GPT-5.4 mini leads
GPT-5.4 mini: 74.1 (#102), Qwen3 235B-A22B: 72.6 (#136)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Instruction Following | 1405 | 1408 |
| IFEval | — | 83.5% |
Long Context Qwen3 235B-A22B leads
GPT-5.4 mini: 43.0 (#112), Qwen3 235B-A22B: 46.1 (#26)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1407 | 1426 |
| Fiction.LiveBench | — | 75% |
Writing & Preference GPT-5.4 mini leads
GPT-5.4 mini: 64.0 (#58), Qwen3 235B-A22B: 59.6 (#108)
| Benchmark | GPT-5.4 mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1412 | 1419 |
| LMArena Creative Writing | 1370 | 1384 |
| EQ-Bench Creative Writing | 1665 | 1366 |
| LMArena Multi-Turn | 1429 | 1432 |
| Short-Story Creative Writing | — | 83% |
| WildBench | — | 86.6% |
Frequently asked questions
Is GPT-5.4 mini better than Qwen3 235B-A22B?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 43.5 on the Noometry Index.
Which is cheaper, GPT-5.4 mini or Qwen3 235B-A22B?
Qwen3 235B-A22B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.
Is GPT-5.4 mini or Qwen3 235B-A22B better for coding?
They score almost the same on coding (45.2 vs 44.3); test both on your own repository before choosing.
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 Qwen3 235B-A22B share?
36 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Qwen3 235B-A22B has 49.