# Gemini 2.5 Pro vs GPT-4o

> Gemini 2.5 Pro is the stronger model overall, scoring 45.0 to 28.6 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gemini-2-5-pro-vs-gpt-4o
- Last updated: 2026-10-10
- Shared benchmarks: 63

## Summary

- They share 63 benchmarks with published results for both. Gemini 2.5 Pro scores higher in 10 categories and GPT-4o in 0 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Gemini 2.5 Pro leads 56.0 to 28.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 84.7% for Gemini 2.5 Pro and 6.4% for GPT-4o.
- Gemini 2.5 Pro is cheaper at $1.25 / $10 per million input/output tokens, against $2.50 / $10 for GPT-4o.
- Gemini 2.5 Pro accepts more context: 1.05M tokens versus 128K.

## Snapshot

| | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 45.0 | 28.6 |
| Rank | 75 | 324 |
| Context | 1.05M | 128K |
| Input $/M | $1.25 | $2.50 |
| Output $/M | $10 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 2.5 Pro: 42.4 (#101)
- GPT-4o: 24.8 (#328)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| SWE-bench Verified | 57.6% | 31% |
| SWE-bench Verified (bash only) | 53.6% | 21.6% |
| Aider Polyglot | 83.1% | 45.3% |
| GSO | 3.9% | 0% |
| WeirdML | 54% | 25.1% |
| LiveBench Coding | 85.9% | 51.4% |
| LMArena Coding | 1452 | 1297 |
| CadEval | 64% | 26% |
| LMArena WebDev | 1227 | — |
| SciCode | 42.8% | — |
| BigCodeBench Instruct | — | 51.1% |
| BigCodeBench Complete | — | 61.1% |
| ALE-Bench | 785.52 | — |
| AlgoTune | 1.51 | — |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 72.2% |

## Agentic & Tool Use

- Gemini 2.5 Pro: 29.2 (#88)
- GPT-4o: 21.0 (#141)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| GDPval | 23.3% | 9.9% |
| TheAgentCompany | 30.3% | 8.6% |
| BALROG | 43.3% | 32.3% |
| LMArena Search | 1142 | 1006 |
| METR Time Horizons | 55.4% | 40.8% |
| Terminal-Bench | 32.6% | — |
| Remote Labor Index | 0.8% | — |
| τ²-bench Banking | 13.7% | — |
| Cybench | — | 12.5% |
| DeepResearch Bench | 42.8% | — |
| Vending-Bench 2 | 573.64 | — |

## Reasoning

- Gemini 2.5 Pro: 28.8 (#99)
- GPT-4o: 9.4 (#343)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| ARC-AGI-2 | 4.9% | 0% |
| SimpleBench | 62.4% | 17.8% |
| ARC-AGI-1 | 41% | 4.5% |
| CritPt | 2% | 0% |
| Chess Puzzles | 20% | 13% |
| EnigmaEval | 5.6% | 0.8% |
| LiveBench Reasoning | 89.8% | 55.8% |
| LMArena Hard Prompts | 1455 | 1281 |
| DTBench | 82.4% | 64.5% |
| LiveBench Data Analysis | 79.9% | 60.9% |
| LMCA | 34.8% | 16.6% |
| Epoch Capabilities Index | 145.32 | 128.97 |
| ForecastBench | 61.3 | 57.7 |
| LiveBench | 82.3% | 55.3% |
| Kagi LLM Benchmark | 70.3% | — |

## Math

- Gemini 2.5 Pro: 32.5 (#213)
- GPT-4o: 10.6 (#312)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| FrontierMath (Tiers 1-3) | 24.6% | 0.4% |
| OTIS Mock AIME 2024-2025 | 84.7% | 6.4% |
| Omni-MATH | 41.6% | 29.3% |
| LiveBench Math | 90.2% | 49.5% |
| LMArena Math | 1450 | 1285 |
| MATH Level 5 | 95.9% | 53.3% |
| FrontierMath (Feb 2025 set) | 14.1% | 0.3% |
| FrontierMath Tier 4 | 0% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |

## Knowledge

- Gemini 2.5 Pro: 56.0 (#46)
- GPT-4o: 28.8 (#242)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| GPQA Diamond | 85.3% | 49.2% |
| Humanity's Last Exam | 21.6% | 2.7% |
| MMLU-Pro | 86.3% | 71.3% |
| Confabulations | 10.6% | 15.3% |
| Vectara Hallucination Rate | 7% | 9.6% |
| GPQA (HELM) | 74.9% | 52% |
| LMArena Expert | 1452 | 1250 |
| SimpleQA Verified | — | 26% |
| MMLU | — | 88.1% |

## Multimodal

- Gemini 2.5 Pro: 45.2 (#18)
- GPT-4o: 34.5 (#91)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| LMArena Vision | 1263 | 1137 |
| GeoBench | 86% | 71% |
| VPCT | 48% | 40% |
| Video-MME | — | 71.9% |
| LMArena Document | 1421 | — |
| ScienceQA | — | 88.5% |
| SpatialViz-Bench | 44.7% | — |

## Multilingual

- Gemini 2.5 Pro: 55.3 (#31)
- GPT-4o: 43.2 (#186)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| LMArena Non-English | 1451 | 1283 |
| LMArena Chinese | 1507 | 1277 |
| LMArena French | 1472 | 1304 |
| LMArena German | 1487 | 1282 |
| LMArena Japanese | 1461 | 1257 |
| LMArena Korean | 1434 | 1234 |
| LMArena Russian | 1461 | 1286 |
| LMArena Spanish | 1473 | 1292 |

## Instruction Following

- Gemini 2.5 Pro: 75.0 (#75)
- GPT-4o: 66.6 (#207)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| LiveBench Instruction Following | 80.6% | 68.6% |
| IFEval | 84% | 81.7% |
| LMArena Instruction Following | 1437 | 1278 |

## Long Context

- Gemini 2.5 Pro: 59.8 (#5)
- GPT-4o: 39.4 (#179)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| Fiction.LiveBench | 91.7% | 66.7% |
| LMArena Longer Query | 1449 | 1289 |

## Writing & Preference

- Gemini 2.5 Pro: 63.7 (#62)
- GPT-4o: 52.6 (#166)

| Benchmark | Gemini 2.5 Pro | GPT-4o |
|---|---|---|
| LMArena Text | 1458 | 1300 |
| LMArena Creative Writing | 1454 | 1292 |
| Short-Story Creative Writing | 83.8% | 81.8% |
| WildBench | 85.7% | 82.8% |
| LMArena Multi-Turn | 1453 | 1302 |
| LiveBench Language | 67.8% | 47.6% |
| EQ-Bench Creative Writing | 1421 | — |

## FAQ

### Is Gemini 2.5 Pro better than GPT-4o?

Gemini 2.5 Pro is the stronger model overall, scoring 45.0 to 28.6 on the Noometry Index.

### Which is cheaper, Gemini 2.5 Pro or GPT-4o?

Gemini 2.5 Pro is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; GPT-4o lists at $2.50 and $10.

### Is Gemini 2.5 Pro or GPT-4o better for coding?

Gemini 2.5 Pro scores higher on coding benchmarks: 42.4 versus 24.8 in the Noometry coding category.

### Which has the bigger context window?

Gemini 2.5 Pro does, with 1.05M tokens against 128K.

### How many benchmarks do Gemini 2.5 Pro and GPT-4o share?

63 benchmarks have published results for both models. Gemini 2.5 Pro has 78 scored results on Noometry and GPT-4o has 72.