Model comparison
Gemini 2.5 Pro vs GPT-5
GPT-5 is the stronger model overall, scoring 50.9 to 45.0 on the Noometry Index.
Last verified . 65 shared benchmarks.
GoogleSummary
- They share 65 benchmarks with published results for both. Gemini 2.5 Pro scores higher in 3 categories and GPT-5 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 leads 55.0 to 32.5.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 24.6% for Gemini 2.5 Pro and 55.4% for GPT-5.
- Both cost about the same: $1.25 input and $10 output per million tokens.
- Gemini 2.5 Pro accepts more context: 1.05M tokens versus 400K.
Side by side
| Gemini 2.5 Pro | GPT-5 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 45.0 | 50.9 |
| Released | 2025-03-25 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 400K |
| Max output | 66K | 128K |
| Input $ / M tokens | $1.25 | $1.25 |
| Output $ / M tokens | $10 | $10 |
| Results tracked | 78 | 69 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
Gemini 2.5 Pro: 42.4 (#101), GPT-5: 50.3 (#47)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| SWE-bench Verified | 57.6% | 73.6% |
| SWE-bench Verified (bash only) | 53.6% | 65% |
| Aider Polyglot | 83.1% | 88% |
| LMArena WebDev | 1227 | 1418 |
| SciCode | 42.8% | 42.9% |
| GSO | 3.9% | 6.9% |
| WeirdML | 54% | 60.7% |
| LMArena Coding | 1452 | 1436 |
| ALE-Bench | 785.52 | 1,162 |
| AlgoTune | 1.51 | 1.67 |
| LiveBench Coding | 85.9% | — |
| CadEval | 64% | — |
Agentic & Tool Use GPT-5 leads
Gemini 2.5 Pro: 29.2 (#88), GPT-5: 33.1 (#56)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| Terminal-Bench | 32.6% | 49.6% |
| GDPval | 23.3% | 34.8% |
| Remote Labor Index | 0.8% | 1.7% |
| DeepResearch Bench | 42.8% | 49.6% |
| BALROG | 43.3% | 32.8% |
| LMArena Search | 1142 | 1133 |
| METR Time Horizons | 55.4% | 69.6% |
| TheAgentCompany | 30.3% | — |
| τ²-bench Banking | 13.7% | — |
| Vending-Bench 2 | 573.64 | — |
Reasoning GPT-5 leads
Gemini 2.5 Pro: 28.8 (#99), GPT-5: 38.3 (#64)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 4.9% | 9.9% |
| SimpleBench | 62.4% | 56.7% |
| Kagi LLM Benchmark | 70.3% | 72.7% |
| ARC-AGI-1 | 41% | 65.7% |
| CritPt | 2% | 12.6% |
| Chess Puzzles | 20% | 37% |
| EnigmaEval | 5.6% | 10.5% |
| LMArena Hard Prompts | 1455 | 1416 |
| DTBench | 82.4% | 90.7% |
| LMCA | 34.8% | 40% |
| Epoch Capabilities Index | 145.32 | 150 |
| ForecastBench | 61.3 | 61.4 |
| EBR-Bench | — | 12.7% |
| LiveBench Reasoning | 89.8% | — |
| Mystery Game Puzzles | — | 23% |
| LiveBench Data Analysis | 79.9% | — |
| LiveBench | 82.3% | — |
Math GPT-5 leads
Gemini 2.5 Pro: 32.5 (#213), GPT-5: 55.0 (#44)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 24.6% | 55.4% |
| FrontierMath Tier 4 | 0% | 22% |
| OTIS Mock AIME 2024-2025 | 84.7% | 91.4% |
| Omni-MATH | 41.6% | 64.7% |
| LMArena Math | 1450 | 1407 |
| MATH Level 5 | 95.9% | 98.1% |
| FrontierMath (Feb 2025 set) | 14.1% | 32.4% |
| FrontierMath Tier 4 (v1) | 4.2% | 12.5% |
| ProofBench | — | 18% |
| LiveBench Math | 90.2% | — |
Knowledge Too close to call
Gemini 2.5 Pro: 56.0 (#46), GPT-5: 56.6 (#43)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| GPQA Diamond | 85.3% | 86.2% |
| Humanity's Last Exam | 21.6% | 25.3% |
| MMLU-Pro | 86.3% | 86.3% |
| Confabulations | 10.6% | 10.3% |
| Vectara Hallucination Rate | 7% | 14.7% |
| GPQA (HELM) | 74.9% | 79.2% |
| LMArena Expert | 1452 | 1419 |
| SimpleQA Verified | — | 50.1% |
Multimodal GPT-5 leads
Gemini 2.5 Pro: 45.2 (#18), GPT-5: 46.8 (#13)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| LMArena Vision | 1263 | 1232 |
| GeoBench | 86% | 81% |
| VPCT | 48% | 66% |
| LMArena Document | 1421 | — |
| SpatialViz-Bench | 44.7% | — |
Multilingual Gemini 2.5 Pro leads
Gemini 2.5 Pro: 55.3 (#31), GPT-5: 51.4 (#110)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| LMArena Non-English | 1451 | 1397 |
| LMArena Chinese | 1507 | 1422 |
| LMArena French | 1472 | 1410 |
| LMArena German | 1487 | 1416 |
| LMArena Japanese | 1461 | 1409 |
| LMArena Korean | 1434 | 1360 |
| LMArena Russian | 1461 | 1406 |
| LMArena Spanish | 1473 | 1399 |
Instruction Following Gemini 2.5 Pro leads
Gemini 2.5 Pro: 75.0 (#75), GPT-5: 73.8 (#113)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| IFEval | 84% | 87.5% |
| LMArena Instruction Following | 1437 | 1388 |
| LiveBench Instruction Following | 80.6% | — |
Long Context GPT-5 leads
Gemini 2.5 Pro: 59.8 (#5), GPT-5: 69.5 (#2)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| Fiction.LiveBench | 91.7% | 97.2% |
| LMArena Longer Query | 1449 | 1399 |
Writing & Preference Too close to call
Gemini 2.5 Pro: 63.7 (#62), GPT-5: 63.4 (#65)
| Benchmark | Gemini 2.5 Pro | GPT-5 |
|---|---|---|
| LMArena Text | 1458 | 1406 |
| LMArena Creative Writing | 1454 | 1365 |
| Short-Story Creative Writing | 83.8% | 86% |
| EQ-Bench Creative Writing | 1421 | 1627 |
| WildBench | 85.7% | 85.7% |
| LMArena Multi-Turn | 1453 | 1426 |
| LiveBench Language | 67.8% | — |
Frequently asked questions
Is Gemini 2.5 Pro better than GPT-5?
GPT-5 is the stronger model overall, scoring 50.9 to 45.0 on the Noometry Index.
Which is cheaper, Gemini 2.5 Pro or GPT-5?
GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Gemini 2.5 Pro lists at $1.25 and $10.
Is Gemini 2.5 Pro or GPT-5 better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 42.4 in the Noometry coding category.
Which has the bigger context window?
Gemini 2.5 Pro does, with 1.05M tokens against 400K.
How many benchmarks do Gemini 2.5 Pro and GPT-5 share?
65 benchmarks have published results for both models. Gemini 2.5 Pro has 78 scored results on Noometry and GPT-5 has 69.