# DeepSeek-R1 vs Gemini 2.5 Pro

> Gemini 2.5 Pro is the stronger model overall, scoring 45.0 to 42.3 on the Noometry Index. DeepSeek-R1 costs 3.8× less per token, which makes it the better buy when Gemini 2.5 Pro's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/deepseek-r1-vs-gemini-2-5-pro
- Last updated: 2026-10-10
- Shared benchmarks: 52

## Summary

- They share 52 benchmarks with published results for both. DeepSeek-R1 scores higher in 3 categories and Gemini 2.5 Pro in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where Gemini 2.5 Pro leads 59.8 to 45.4.
- The biggest single-benchmark swing is SimpleBench: 40.8% for DeepSeek-R1 and 62.4% for Gemini 2.5 Pro.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $1.25 / $10 for Gemini 2.5 Pro.
- Gemini 2.5 Pro accepts more context: 1.05M tokens versus 164K.

## Snapshot

| | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| Provider | DeepSeek | Google |
| Noometry Index | 42.3 | 45.0 |
| Rank | 115 | 75 |
| Context | 164K | 1.05M |
| Input $/M | $0.50 | $1.25 |
| Output $/M | $2.15 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- DeepSeek-R1: 46.3 (#68)
- Gemini 2.5 Pro: 42.4 (#101)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| Aider Polyglot | 71.4% | 83.1% |
| SciCode | 35.7% | 42.8% |
| WeirdML | 41.6% | 54% |
| LiveBench Coding | 66.7% | 85.9% |
| LMArena Coding | 1427 | 1452 |
| ALE-Bench | 804.12 | 785.52 |
| AlgoTune | 1.7 | 1.51 |
| SWE-bench Verified | — | 57.6% |
| SWE-bench Verified (bash only) | — | 53.6% |
| LMArena WebDev | — | 1227 |
| GSO | — | 3.9% |
| CadEval | — | 64% |

## Agentic & Tool Use

- DeepSeek-R1: 30.7 (#75)
- Gemini 2.5 Pro: 29.2 (#88)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| DeepResearch Bench | 35.1% | 42.8% |
| BALROG | 34.9% | 43.3% |
| METR Time Horizons | 53.8% | 55.4% |
| Terminal-Bench | — | 32.6% |
| GDPval | — | 23.3% |
| Remote Labor Index | — | 0.8% |
| TheAgentCompany | — | 30.3% |
| τ²-bench Banking | — | 13.7% |
| LMArena Search | — | 1142 |
| Vending-Bench 2 | — | 573.64 |

## Reasoning

- DeepSeek-R1: 18.6 (#278)
- Gemini 2.5 Pro: 28.8 (#99)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| ARC-AGI-2 | 1.3% | 4.9% |
| SimpleBench | 40.8% | 62.4% |
| Kagi LLM Benchmark | 69.4% | 70.3% |
| ARC-AGI-1 | 21.2% | 41% |
| CritPt | 1.1% | 2% |
| LiveBench Reasoning | 83.2% | 89.8% |
| LMArena Hard Prompts | 1416 | 1455 |
| LiveBench Data Analysis | 69.8% | 79.9% |
| Epoch Capabilities Index | 141.29 | 145.32 |
| ForecastBench | 60 | 61.3 |
| LiveBench | 71.6% | 82.3% |
| Chess Puzzles | — | 20% |
| EnigmaEval | — | 5.6% |
| DTBench | — | 82.4% |
| LMCA | — | 34.8% |

## Math

- DeepSeek-R1: 43.8 (#79)
- Gemini 2.5 Pro: 32.5 (#213)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 84.7% |
| Omni-MATH | 42.4% | 41.6% |
| LiveBench Math | 80.7% | 90.2% |
| LMArena Math | 1400 | 1450 |
| MATH Level 5 | 96.6% | 95.9% |
| FrontierMath (Tiers 1-3) | — | 24.6% |
| FrontierMath Tier 4 | — | 0% |
| FrontierMath (Feb 2025 set) | — | 14.1% |
| FrontierMath Tier 4 (v1) | — | 4.2% |

## Knowledge

- DeepSeek-R1: 44.5 (#87)
- Gemini 2.5 Pro: 56.0 (#46)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| GPQA Diamond | 76.3% | 85.3% |
| MMLU-Pro | 79.3% | 86.3% |
| Confabulations | 12.7% | 10.6% |
| Vectara Hallucination Rate | 11.3% | 7% |
| GPQA (HELM) | 66.6% | 74.9% |
| LMArena Expert | 1394 | 1452 |
| Humanity's Last Exam | — | 21.6% |

## Multimodal

- DeepSeek-R1: —
- Gemini 2.5 Pro: 45.2 (#18)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Vision | — | 1263 |
| GeoBench | — | 86% |
| VPCT | — | 48% |
| LMArena Document | — | 1421 |
| SpatialViz-Bench | — | 44.7% |

## Multilingual

- DeepSeek-R1: 52.4 (#85)
- Gemini 2.5 Pro: 55.3 (#31)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Non-English | 1412 | 1451 |
| LMArena Chinese | 1442 | 1507 |
| LMArena French | 1417 | 1472 |
| LMArena German | 1404 | 1487 |
| LMArena Japanese | 1391 | 1461 |
| LMArena Korean | 1360 | 1434 |
| LMArena Russian | 1423 | 1461 |
| LMArena Spanish | 1411 | 1473 |

## Instruction Following

- DeepSeek-R1: 72.0 (#143)
- Gemini 2.5 Pro: 75.0 (#75)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 80.6% |
| IFEval | 78.4% | 84% |
| LMArena Instruction Following | 1382 | 1437 |

## Long Context

- DeepSeek-R1: 45.4 (#36)
- Gemini 2.5 Pro: 59.8 (#5)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| Fiction.LiveBench | 75% | 91.7% |
| LMArena Longer Query | 1391 | 1449 |

## Writing & Preference

- DeepSeek-R1: 61.4 (#88)
- Gemini 2.5 Pro: 63.7 (#62)

| Benchmark | DeepSeek-R1 | Gemini 2.5 Pro |
|---|---|---|
| LMArena Text | 1428 | 1458 |
| LMArena Creative Writing | 1405 | 1454 |
| Short-Story Creative Writing | 83% | 83.8% |
| EQ-Bench Creative Writing | 1500 | 1421 |
| WildBench | 82.8% | 85.7% |
| LMArena Multi-Turn | 1405 | 1453 |
| LiveBench Language | 48.5% | 67.8% |

## FAQ

### Is DeepSeek-R1 better than Gemini 2.5 Pro?

Gemini 2.5 Pro is the stronger model overall, scoring 45.0 to 42.3 on the Noometry Index. DeepSeek-R1 costs 3.8× less per token, which makes it the better buy when Gemini 2.5 Pro's lead doesn't matter for your workload.

### Which is cheaper, DeepSeek-R1 or Gemini 2.5 Pro?

DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Gemini 2.5 Pro lists at $1.25 and $10.

### Is DeepSeek-R1 or Gemini 2.5 Pro better for coding?

DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 42.4 in the Noometry coding category.

### Which has the bigger context window?

Gemini 2.5 Pro does, with 1.05M tokens against 164K.

### How many benchmarks do DeepSeek-R1 and Gemini 2.5 Pro share?

52 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Gemini 2.5 Pro has 78.