Model comparison
Gemini 3 Pro vs GPT-5
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 50.9 on the Noometry Index.
Last verified . 56 shared benchmarks.
GoogleSummary
- They share 56 benchmarks with published results for both. Gemini 3 Pro scores higher in 8 categories and GPT-5 in 2 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 44.0.
- The biggest single-benchmark swing is BALROG: 58.1% for Gemini 3 Pro and 32.8% for GPT-5.
Side by side
| Gemini 3 Pro | GPT-5 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 54.8 | 50.9 |
| Released | 2025-11-18 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | — | 400K |
| Max output | — | 128K |
| Input $ / M tokens | — | $1.25 |
| Output $ / M tokens | — | $10 |
| Results tracked | 67 | 69 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Pro leads
Gemini 3 Pro: 51.6 (#39), GPT-5: 50.3 (#47)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| SWE-bench Verified | 72.9% | 73.6% |
| SWE-bench Verified (bash only) | 74.2% | 65% |
| LMArena WebDev | 1440 | 1418 |
| GSO | 18.6% | 6.9% |
| WeirdML | 69.9% | 60.7% |
| LMArena Coding | 1481 | 1436 |
| ALE-Bench | 1,177 | 1,162 |
| AlgoTune | 1.83 | 1.67 |
| Aider Polyglot | — | 88% |
| SWE-bench Multilingual | 68.7% | — |
| SciCode | — | 42.9% |
Agentic & Tool Use Gemini 3 Pro leads
Gemini 3 Pro: 40.6 (#23), GPT-5: 33.1 (#56)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| Terminal-Bench | 69.4% | 49.6% |
| GDPval | 40.3% | 34.8% |
| Remote Labor Index | 1.3% | 1.7% |
| DeepResearch Bench | 46.3% | 49.6% |
| BALROG | 58.1% | 32.8% |
| LMArena Search | 1207 | 1133 |
| METR Time Horizons | 71% | 69.6% |
| Berkeley Function Calling Leaderboard | 72.5% | — |
| τ²-bench Airline | 80.5% | — |
| τ²-bench Banking | 18% | — |
| τ²-bench Retail | 75.9% | — |
| τ²-bench Telecom | 91% | — |
| Vending-Bench 2 | 5,478 | — |
Reasoning Gemini 3 Pro leads
Gemini 3 Pro: 52.5 (#31), GPT-5: 38.3 (#64)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 31.1% | 9.9% |
| SimpleBench | 76.4% | 56.7% |
| Kagi LLM Benchmark | 80.1% | 72.7% |
| ARC-AGI-1 | 75% | 65.7% |
| CritPt | 6.9% | 12.6% |
| Chess Puzzles | 31% | 37% |
| EnigmaEval | 18.2% | 10.5% |
| LMArena Hard Prompts | 1480 | 1416 |
| Epoch Capabilities Index | 152.92 | 150 |
| ForecastBench | 61.2 | 61.4 |
| NYT Connections (extended) | 94.4% | — |
| EBR-Bench | — | 12.7% |
| Mystery Game Puzzles | — | 23% |
| DTBench | — | 90.7% |
| LMCA | — | 40% |
Math GPT-5 leads
Gemini 3 Pro: 49.9 (#59), GPT-5: 55.0 (#44)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.4% | 91.4% |
| ProofBench | 20% | 18% |
| Omni-MATH | 55.5% | 64.7% |
| LMArena Math | 1476 | 1407 |
| FrontierMath (Feb 2025 set) | 37.6% | 32.4% |
| FrontierMath Tier 4 (v1) | 18.8% | 12.5% |
| FrontierMath (Tiers 1-3) | — | 55.4% |
| FrontierMath Tier 4 | — | 22% |
| MathArena Final-Answer Competitions | 67% | — |
| MATH Level 5 | — | 98.1% |
Knowledge Gemini 3 Pro leads
Gemini 3 Pro: 64.4 (#16), GPT-5: 56.6 (#43)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| GPQA Diamond | 92.6% | 86.2% |
| Humanity's Last Exam | 37.5% | 25.3% |
| MMLU-Pro | 90.3% | 86.3% |
| Vectara Hallucination Rate | 13.6% | 14.7% |
| GPQA (HELM) | 80.3% | 79.2% |
| LMArena Expert | 1475 | 1419 |
| SimpleQA Verified | — | 50.1% |
| Confabulations | — | 10.3% |
Multimodal Gemini 3 Pro leads
Gemini 3 Pro: 57.6 (#2), GPT-5: 46.8 (#13)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| LMArena Vision | 1305 | 1232 |
| GeoBench | 84% | 81% |
| VPCT | 91% | 66% |
| LMArena Document | 1434 | — |
Multilingual Gemini 3 Pro leads
Gemini 3 Pro: 56.9 (#16), GPT-5: 51.4 (#110)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| LMArena Non-English | 1474 | 1397 |
| LMArena Chinese | 1523 | 1422 |
| LMArena French | 1492 | 1410 |
| LMArena German | 1515 | 1416 |
| LMArena Japanese | 1510 | 1409 |
| LMArena Korean | 1448 | 1360 |
| LMArena Russian | 1493 | 1406 |
| LMArena Spanish | 1470 | 1399 |
Instruction Following Gemini 3 Pro leads
Gemini 3 Pro: 76.3 (#45), GPT-5: 73.8 (#113)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| IFEval | 87.7% | 87.5% |
| LMArena Instruction Following | 1458 | 1388 |
Long Context GPT-5 leads
Gemini 3 Pro: 44.0 (#79), GPT-5: 69.5 (#2)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| LMArena Longer Query | 1471 | 1399 |
| Fiction.LiveBench | — | 97.2% |
| CL-bench | 15.8% | — |
Writing & Preference Gemini 3 Pro leads
Gemini 3 Pro: 66.4 (#35), GPT-5: 63.4 (#65)
| Benchmark | Gemini 3 Pro | GPT-5 |
|---|---|---|
| LMArena Text | 1479 | 1406 |
| LMArena Creative Writing | 1482 | 1365 |
| EQ-Bench Creative Writing | 1525 | 1627 |
| WildBench | 85.9% | 85.7% |
| LMArena Multi-Turn | 1484 | 1426 |
| Short-Story Creative Writing | — | 86% |
Frequently asked questions
Is Gemini 3 Pro better than GPT-5?
Gemini 3 Pro is the stronger model overall, scoring 54.8 to 50.9 on the Noometry Index.
Is Gemini 3 Pro or GPT-5 better for coding?
Gemini 3 Pro scores higher on coding benchmarks: 51.6 versus 50.3 in the Noometry coding category.
How many benchmarks do Gemini 3 Pro and GPT-5 share?
56 benchmarks have published results for both models. Gemini 3 Pro has 67 scored results on Noometry and GPT-5 has 69.