Model comparison
o4-mini vs Qwen3 235B-A22B
Qwen3 235B-A22B is the stronger model overall, scoring 43.5 to 41.6 on the Noometry Index.
Last verified . 46 shared benchmarks.
Alibaba (Qwen)Summary
- They share 46 benchmarks with published results for both. o4-mini scores higher in 2 categories and Qwen3 235B-A22B in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3 235B-A22B leads 50.4 to 40.8.
- The biggest single-benchmark swing is ARC-AGI-1: 58.7% for o4-mini and 11% for Qwen3 235B-A22B.
- Qwen3 235B-A22B is cheaper at $0.70 / $2.80 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- o4-mini accepts more context: 200K tokens versus 131K.
- Qwen3 235B-A22B has downloadable open weights; the other is API-only.
Side by side
| o4-mini | Qwen3 235B-A22B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 41.6 | 43.5 |
| Released | 2025-04-16 | 2025-04 |
| Weights | Proprietary | Open |
| Context window | 200K | 131K |
| Max output | 100K | 16K |
| Input $ / M tokens | $1.10 | $0.70 |
| Output $ / M tokens | $4.40 | $2.80 |
| Results tracked | 60 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 235B-A22B leads
o4-mini: 40.9 (#127), Qwen3 235B-A22B: 44.3 (#75)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| Aider Polyglot | 72% | 59.6% |
| WeirdML | 52.6% | 41% |
| LMArena Coding | 1368 | 1445 |
| SWE-bench Verified (bash only) | 45% | — |
| SciCode | — | 42.4% |
| GSO | 3.6% | — |
| CadEval | 62% | — |
| ALE-Bench | 826.17 | — |
| AlgoTune | 1.72 | — |
Agentic & Tool Use Qwen3 235B-A22B leads
o4-mini: 32.6 (#61), Qwen3 235B-A22B: 33.9 (#51)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 53.2% | 52.1% |
| GDPval | 25.3% | — |
| METR Time Horizons | 63.9% | — |
| Vending-Bench 2 | — | -11.34 |
Reasoning o4-mini leads
o4-mini: 24.6 (#162), Qwen3 235B-A22B: 15.7 (#311)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| ARC-AGI-2 | 6.1% | 1.3% |
| SimpleBench | 38.7% | 31% |
| Kagi LLM Benchmark | 67.6% | 69.4% |
| ARC-AGI-1 | 58.7% | 11% |
| CritPt | 0.6% | 0% |
| Chess Puzzles | 26% | 12% |
| LMArena Hard Prompts | 1351 | 1433 |
| Mystery Game Puzzles | 5% | 9% |
| DTBench | 77.6% | 80.3% |
| LMCA | 26.5% | 29.3% |
| Epoch Capabilities Index | 145.64 | 143.85 |
| ForecastBench | 61.8 | 59.7 |
| EnigmaEval | 9.2% | — |
Math Qwen3 235B-A22B leads
o4-mini: 40.8 (#89), Qwen3 235B-A22B: 50.4 (#57)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 81.7% | 86.7% |
| Omni-MATH | 72% | 71.8% |
| LMArena Math | 1389 | 1432 |
| MATH Level 5 | 97.8% | 68.9% |
| FrontierMath (Feb 2025 set) | 24.8% | 8.5% |
| FrontierMath Tier 4 (v1) | 6.3% | 0% |
| FrontierMath (Tiers 1-3) | 36.1% | — |
| FrontierMath Tier 4 | 4.9% | — |
Knowledge Qwen3 235B-A22B leads
o4-mini: 43.6 (#91), Qwen3 235B-A22B: 49.6 (#73)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| GPQA Diamond | 79.6% | 80.1% |
| SimpleQA Verified | 19.6% | 40.4% |
| MMLU-Pro | 82% | 84.4% |
| Confabulations | 15.8% | 15.6% |
| Vectara Hallucination Rate | 18.6% | 9.3% |
| GPQA (HELM) | 73.5% | 72.7% |
| LMArena Expert | 1343 | 1463 |
| Humanity's Last Exam | 18.1% | — |
Multimodal Not comparable
o4-mini: 40.2 (#49), Qwen3 235B-A22B: —
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Vision | 1194 | — |
| GeoBench | 64% | — |
| VPCT | 57.5% | — |
Multilingual Qwen3 235B-A22B leads
o4-mini: 47.0 (#154), Qwen3 235B-A22B: 52.3 (#89)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1337 | 1409 |
| LMArena Chinese | 1354 | 1481 |
| LMArena French | 1364 | 1445 |
| LMArena German | 1336 | 1433 |
| LMArena Japanese | 1308 | 1399 |
| LMArena Korean | 1312 | 1391 |
| LMArena Russian | 1334 | 1411 |
| LMArena Spanish | 1347 | 1430 |
Instruction Following o4-mini leads
o4-mini: 75.2 (#68), Qwen3 235B-A22B: 72.6 (#136)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| IFEval | 92.8% | 83.5% |
| LMArena Instruction Following | 1321 | 1408 |
Long Context Too close to call
o4-mini: 45.5 (#33), Qwen3 235B-A22B: 46.1 (#26)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| Fiction.LiveBench | 77.8% | 75% |
| LMArena Longer Query | 1315 | 1426 |
Writing & Preference Qwen3 235B-A22B leads
o4-mini: 54.0 (#152), Qwen3 235B-A22B: 59.6 (#108)
| Benchmark | o4-mini | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1353 | 1419 |
| LMArena Creative Writing | 1294 | 1384 |
| Short-Story Creative Writing | 75% | 83% |
| WildBench | 85.4% | 86.6% |
| LMArena Multi-Turn | 1350 | 1432 |
| EQ-Bench Creative Writing | — | 1366 |
Frequently asked questions
Is o4-mini better than Qwen3 235B-A22B?
Qwen3 235B-A22B is the stronger model overall, scoring 43.5 to 41.6 on the Noometry Index.
Which is cheaper, o4-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; o4-mini lists at $1.10 and $4.40.
Is o4-mini or Qwen3 235B-A22B better for coding?
Qwen3 235B-A22B scores higher on coding benchmarks: 44.3 versus 40.9 in the Noometry coding category.
Which has the bigger context window?
o4-mini does, with 200K tokens against 131K.
How many benchmarks do o4-mini and Qwen3 235B-A22B share?
46 benchmarks have published results for both models. o4-mini has 60 scored results on Noometry and Qwen3 235B-A22B has 49.