Model comparison
Gemini 3 Pro vs Qwen3 235B-A22B
Gemini 3 Pro is the stronger model overall, scoring 54.7 to 43.2 on the Noometry Index.
Last verified . 39 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 39 benchmarks with published results for both. Gemini 3 Pro scores higher in 8 categories and Qwen3 235B-A22B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3 Pro leads 52.5 to 15.8.
- The biggest single-benchmark swing is ARC-AGI-1: 75% for Gemini 3 Pro and 11% for Qwen3 235B-A22B.
- Qwen3 235B-A22B has downloadable open weights; the other is API-only.
Side by side
| Gemini 3 Pro | Qwen3 235B-A22B | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 54.7 | 43.2 |
| Released | 2025-11-18 | 2025-04 |
| Weights | Proprietary | Open |
| Context window | — | 131K |
| Max output | — | 16K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $2.80 |
| Results tracked | 67 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Pro leads
Gemini 3 Pro: 52.1 (#33), Qwen3 235B-A22B: 43.8 (#72)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| WeirdML | 69.9% | 41% |
| LMArena Coding | 1482 | 1445 |
| SWE-bench Verified | 72.9% | — |
| SWE-bench Verified (bash only) | 74.2% | — |
| Aider Polyglot | — | 59.6% |
| LMArena WebDev | 1439 | — |
| SWE-bench Multilingual | 68.7% | — |
| SciCode | — | 42.4% |
| GSO | 18.6% | — |
| ALE-Bench | 1,177 | — |
| AlgoTune | 1.83 | — |
Agentic & Tool Use Gemini 3 Pro leads
Gemini 3 Pro: 40.5 (#21), Qwen3 235B-A22B: 33.7 (#48)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 72.5% | 52.1% |
| Vending-Bench 2 | 5,478 | -11.34 |
| Terminal-Bench | 69.4% | — |
| GDPval | 40.3% | — |
| Remote Labor Index | 1.3% | — |
| τ²-bench Airline | 80.5% | — |
| τ²-bench Banking | 18% | — |
| τ²-bench Retail | 75.9% | — |
| τ²-bench Telecom | 91% | — |
| DeepResearch Bench | 46.3% | — |
| BALROG | 58.1% | — |
| LMArena Search | 1207 | — |
| METR Time Horizons | 71% | — |
Reasoning Gemini 3 Pro leads
Gemini 3 Pro: 52.5 (#28), Qwen3 235B-A22B: 15.8 (#302)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| ARC-AGI-2 | 31.1% | 1.3% |
| SimpleBench | 76.4% | 31% |
| Kagi LLM Benchmark | 80.1% | 69.4% |
| ARC-AGI-1 | 75% | 11% |
| CritPt | 6.9% | 0% |
| Chess Puzzles | 31% | 12% |
| LMArena Hard Prompts | 1481 | 1433 |
| Epoch Capabilities Index | 152.96 | 143.87 |
| ForecastBench | 61.2 | 59.7 |
| NYT Connections (extended) | 94.4% | — |
| EnigmaEval | 18.2% | — |
| Mystery Game Puzzles | — | 9% |
| DTBench | — | 80.3% |
| LMCA | — | 29.3% |
Math Too close to call
Gemini 3 Pro: 48.8 (#50), Qwen3 235B-A22B: 48.5 (#51)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.4% | 86.7% |
| Omni-MATH | 55.5% | 71.8% |
| LMArena Math | 1475 | 1431 |
| FrontierMath (Feb 2025 set) | 37.6% | 8.5% |
| FrontierMath Tier 4 (v1) | 18.8% | 0% |
| MathArena Final-Answer Competitions | 67% | — |
| ProofBench | 20% | — |
| MATH Level 5 | — | 68.9% |
Knowledge Gemini 3 Pro leads
Gemini 3 Pro: 64.3 (#15), Qwen3 235B-A22B: 49.5 (#67)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| GPQA Diamond | 92.6% | 80.1% |
| MMLU-Pro | 90.3% | 84.4% |
| Vectara Hallucination Rate | 13.6% | 9.3% |
| GPQA (HELM) | 80.3% | 72.7% |
| LMArena Expert | 1475 | 1460 |
| Humanity's Last Exam | 37.5% | — |
| SimpleQA Verified | — | 40.4% |
| Confabulations | — | 15.6% |
Multimodal Not comparable
Gemini 3 Pro: 56.9 (#1), Qwen3 235B-A22B: —
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| LMArena Vision | 1305 | — |
| GeoBench | 84% | — |
| VPCT | 91% | — |
| LMArena Document | 1434 | — |
Multilingual Gemini 3 Pro leads
Gemini 3 Pro: 57.1 (#14), Qwen3 235B-A22B: 52.5 (#84)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1474 | 1410 |
| LMArena Chinese | 1522 | 1481 |
| LMArena French | 1492 | 1445 |
| LMArena German | 1517 | 1432 |
| LMArena Japanese | 1513 | 1400 |
| LMArena Korean | 1449 | 1391 |
| LMArena Russian | 1493 | 1411 |
| LMArena Spanish | 1468 | 1429 |
Instruction Following Gemini 3 Pro leads
Gemini 3 Pro: 76.4 (#42), Qwen3 235B-A22B: 72.8 (#132)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| IFEval | 87.7% | 83.5% |
| LMArena Instruction Following | 1458 | 1408 |
Long Context Qwen3 235B-A22B leads
Gemini 3 Pro: 44.0 (#75), Qwen3 235B-A22B: 46.1 (#25)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1472 | 1426 |
| Fiction.LiveBench | — | 75% |
| CL-bench | 15.8% | — |
Writing & Preference Gemini 3 Pro leads
Gemini 3 Pro: 66.6 (#34), Qwen3 235B-A22B: 59.8 (#103)
| Benchmark | Gemini 3 Pro | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1479 | 1419 |
| LMArena Creative Writing | 1482 | 1384 |
| EQ-Bench Creative Writing | 1525 | 1366 |
| WildBench | 85.9% | 86.6% |
| LMArena Multi-Turn | 1484 | 1432 |
| Short-Story Creative Writing | — | 83% |
Frequently asked questions
Is Gemini 3 Pro better than Qwen3 235B-A22B?
Gemini 3 Pro is the stronger model overall, scoring 54.7 to 43.2 on the Noometry Index.
Is Gemini 3 Pro or Qwen3 235B-A22B better for coding?
Gemini 3 Pro scores higher on coding benchmarks: 52.1 versus 43.8 in the Noometry coding category.
How many benchmarks do Gemini 3 Pro and Qwen3 235B-A22B share?
39 benchmarks have published results for both models. Gemini 3 Pro has 67 scored results on Noometry and Qwen3 235B-A22B has 49.