Model comparison
Gemini 3.1 Pro Preview vs GPT-6 Sol
GPT-6 Sol is the stronger model overall, scoring 61.8 to 56.7 on the Noometry Index.
Last verified . 44 shared benchmarks.
GoogleSummary
- They share 44 benchmarks with published results for both. Gemini 3.1 Pro Preview scores higher in 5 categories and GPT-6 Sol in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Sol leads 87.2 to 62.1.
- The biggest single-benchmark swing is FrontierMath Tier 4: 26.8% for Gemini 3.1 Pro Preview and 90% for GPT-6 Sol.
- GPT-6 Sol is cheaper at $2 / $10 per million input/output tokens, against $2 / $12 for Gemini 3.1 Pro Preview.
- GPT-6 Sol accepts more context: 1.05M tokens versus 1.05M.
Side by side
| Gemini 3.1 Pro Preview | GPT-6 Sol | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 56.7 | 61.8 |
| Released | 2026-02-19 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 66K | 128K |
| Input $ / M tokens | $2 | $2 |
| Output $ / M tokens | $12 | $10 |
| Results tracked | 71 | 45 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Sol leads
Gemini 3.1 Pro Preview: 42.5 (#99), GPT-6 Sol: 60.1 (#11)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| DeepSWE | 11.7% | 68.8% |
| LMArena WebDev | 1447 | 1688 |
| SciCode | 58.9% | 57.6% |
| LMArena Coding | 1484 | 1447 |
| ALE-Bench | 1,161 | 2,462 |
| SWE-bench Verified | 75.6% | — |
| FrontierCode | — | 49.3% |
| GSO | 22.6% | — |
| WeirdML | 72.1% | — |
| MirrorCode | 8.9% | — |
| AlgoTune | 2.02 | — |
Agentic & Tool Use Too close to call
Gemini 3.1 Pro Preview: 37.7 (#34), GPT-6 Sol: 37.2 (#36)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| APEX-Agents | 35.3% | 54.3% |
| GDP.pdf | 17% | 26.4% |
| Vending-Bench 2 | 3,774 | 14,428 |
| Terminal-Bench | 80.2% | — |
| τ²-bench Banking | 26% | — |
| DeepResearch Bench | 47.8% | — |
| PostTrainBench | 22% | — |
| BALROG | 57% | — |
| ExploitBench | 26.1% | — |
| GBAEval | 0.8% | — |
| LMArena Search | 1211 | — |
| METR Time Horizons | 77% | — |
Reasoning GPT-6 Sol leads
Gemini 3.1 Pro Preview: 71.7 (#12), GPT-6 Sol: 74.0 (#9)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| ARC-AGI-2 | 77.1% | 89.6% |
| NYT Connections (extended) | 97.4% | 90.1% |
| ARC-AGI-1 | 98% | 95.5% |
| CritPt | 17.7% | 30.9% |
| EBR-Bench | 14.3% | 53.3% |
| LMArena Hard Prompts | 1485 | 1418 |
| Mystery Game Puzzles | 34% | 56% |
| DTBench | 97.1% | 97.3% |
| LMCA | 53.8% | 59.1% |
| Epoch Capabilities Index | 154.77 | 162.72 |
| SimpleBench | 79.6% | — |
| Chess Puzzles | 55% | — |
| EnigmaEval | 36.8% | — |
| Thematic Generalization | 79.4% | — |
| ForecastBench | 59 | — |
Math GPT-6 Sol leads
Gemini 3.1 Pro Preview: 62.1 (#34), GPT-6 Sol: 87.2 (#7)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 59.6% | 89.8% |
| FrontierMath Tier 4 | 26.8% | 90% |
| OTIS Mock AIME 2024-2025 | 95.6% | 100% |
| ProofBench | 26% | 83% |
| LMArena Math | 1485 | 1402 |
| MathArena Final-Answer Competitions | 86.5% | — |
| FrontierMath (Feb 2025 set) | 36.9% | — |
| FrontierMath Tier 4 (v1) | 16.7% | — |
Knowledge Gemini 3.1 Pro Preview leads
Gemini 3.1 Pro Preview: 71.8 (#3), GPT-6 Sol: 64.8 (#15)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| GPQA Diamond | 94.4% | 94.3% |
| SimpleQA Verified | 73.5% | 60.7% |
| Vectara Hallucination Rate | 10.4% | 6.5% |
| LMArena Expert | 1485 | 1439 |
| Humanity's Last Exam | 46.4% | — |
Multimodal GPT-6 Sol leads
Gemini 3.1 Pro Preview: 37.9 (#69), GPT-6 Sol: 47.6 (#10)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| LMArena Vision | 1296 | 1245 |
| Blueprint-Bench 2 | 26.5% | 36.9% |
| Furniture Assembly | 26.7% | 58.3% |
| LMArena Document | 1444 | — |
Multilingual Gemini 3.1 Pro Preview leads
Gemini 3.1 Pro Preview: 57.0 (#12), GPT-6 Sol: 50.5 (#118)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| LMArena Non-English | 1477 | 1385 |
| LMArena Chinese | 1529 | 1405 |
| LMArena French | 1487 | 1410 |
| LMArena German | 1491 | 1390 |
| LMArena Japanese | 1493 | 1385 |
| LMArena Korean | 1455 | 1341 |
| LMArena Russian | 1498 | 1401 |
| LMArena Spanish | 1479 | 1384 |
Instruction Following Gemini 3.1 Pro Preview leads
Gemini 3.1 Pro Preview: 77.0 (#32), GPT-6 Sol: 74.5 (#94)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| LMArena Instruction Following | 1466 | 1412 |
Long Context Gemini 3.1 Pro Preview leads
Gemini 3.1 Pro Preview: 47.4 (#18), GPT-6 Sol: 43.1 (#108)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| LMArena Longer Query | 1483 | 1411 |
| CL-bench | 20.8% | — |
| CL-bench Life | 16.9% | — |
Writing & Preference GPT-6 Sol leads
Gemini 3.1 Pro Preview: 66.1 (#37), GPT-6 Sol: 71.9 (#18)
| Benchmark | Gemini 3.1 Pro Preview | GPT-6 Sol |
|---|---|---|
| LMArena Text | 1481 | 1395 |
| LMArena Creative Writing | 1482 | 1378 |
| EQ-Bench Creative Writing | 1491 | 2125 |
| LMArena Multi-Turn | 1488 | 1412 |
| EQ-Bench 4 | 1142 | — |
Frequently asked questions
Is Gemini 3.1 Pro Preview better than GPT-6 Sol?
GPT-6 Sol is the stronger model overall, scoring 61.8 to 56.7 on the Noometry Index.
Which is cheaper, Gemini 3.1 Pro Preview or GPT-6 Sol?
GPT-6 Sol is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.
Is Gemini 3.1 Pro Preview or GPT-6 Sol better for coding?
GPT-6 Sol scores higher on coding benchmarks: 60.1 versus 42.5 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Sol does, with 1.05M tokens against 1.05M.
How many benchmarks do Gemini 3.1 Pro Preview and GPT-6 Sol share?
44 benchmarks have published results for both models. Gemini 3.1 Pro Preview has 71 scored results on Noometry and GPT-6 Sol has 45.