# DeepSeek-R1 vs Gemini 3.5 Flash

> Gemini 3.5 Flash is the stronger model overall, scoring 54.2 to 42.3 on the Noometry Index. DeepSeek-R1 costs 3.7× less per token, which makes it the better buy when Gemini 3.5 Flash's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/deepseek-r1-vs-gemini-3-5-flash
- Last updated: 2026-10-10
- Shared benchmarks: 28

## Summary

- They share 28 benchmarks with published results for both. DeepSeek-R1 scores higher in 2 categories and Gemini 3.5 Flash in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.5 Flash leads 62.8 to 18.6.
- The biggest single-benchmark swing is ARC-AGI-1: 21.2% for DeepSeek-R1 and 92.5% for Gemini 3.5 Flash.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $1.50 / $9 for Gemini 3.5 Flash.
- Gemini 3.5 Flash accepts more context: 1.05M tokens versus 164K.

## Snapshot

| | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| Provider | DeepSeek | Google |
| Noometry Index | 42.3 | 54.2 |
| Rank | 115 | 32 |
| Context | 164K | 1.05M |
| Input $/M | $0.50 | $1.50 |
| Output $/M | $2.15 | $9 |
| Weights | Proprietary | Proprietary |

## Coding

- DeepSeek-R1: 46.3 (#68)
- Gemini 3.5 Flash: 49.4 (#49)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| SciCode | 35.7% | 53.1% |
| WeirdML | 41.6% | 62.6% |
| LMArena Coding | 1427 | 1492 |
| ALE-Bench | 804.12 | 911.02 |
| SWE-bench Verified | — | 79.3% |
| DeepSWE | — | 37.4% |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1499 |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |

## Agentic & Tool Use

- DeepSeek-R1: 30.7 (#75)
- Gemini 3.5 Flash: 24.7 (#114)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| APEX-Agents | — | 27.5% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| GBAEval | — | 6.7% |
| GDP.pdf | — | 14% |
| METR Time Horizons | 53.8% | — |
| Vending-Bench 2 | — | 5,396 |

## Reasoning

- DeepSeek-R1: 18.6 (#278)
- Gemini 3.5 Flash: 62.8 (#18)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| ARC-AGI-2 | 1.3% | 72.1% |
| SimpleBench | 40.8% | 76.7% |
| ARC-AGI-1 | 21.2% | 92.5% |
| CritPt | 1.1% | 13.1% |
| LMArena Hard Prompts | 1416 | 1488 |
| Epoch Capabilities Index | 141.29 | 154.46 |
| ForecastBench | 60 | 59 |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 92.6% |
| Chess Puzzles | — | 50% |
| EnigmaEval | — | 25.4% |
| EBR-Bench | — | 4.8% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 32% |
| DTBench | — | 94.7% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 47.1% |
| Surface Evolver Bench | — | 58.1% |
| LiveBench | 71.6% | — |

## Math

- DeepSeek-R1: 43.8 (#79)
- Gemini 3.5 Flash: 60.7 (#36)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 95.6% |
| LMArena Math | 1400 | 1504 |
| FrontierMath (Tiers 1-3) | — | 62.8% |
| FrontierMath Tier 4 | — | 26.8% |
| MathArena Final-Answer Competitions | — | 76.3% |
| ProofBench | — | 31% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
| FrontierMath (Feb 2025 set) | — | 39% |
| FrontierMath Tier 4 (v1) | — | 14.6% |

## Knowledge

- DeepSeek-R1: 44.5 (#87)
- Gemini 3.5 Flash: 66.3 (#11)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| GPQA Diamond | 76.3% | 92.8% |
| LMArena Expert | 1394 | 1495 |
| SimpleQA Verified | — | 66.2% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |

## Multimodal

- DeepSeek-R1: —
- Gemini 3.5 Flash: 45.7 (#15)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| LMArena Vision | — | 1310 |
| Blueprint-Bench 2 | — | 33.6% |
| LMArena Document | — | 1463 |

## Multilingual

- DeepSeek-R1: 52.4 (#85)
- Gemini 3.5 Flash: 57.0 (#13)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| LMArena Non-English | 1412 | 1476 |
| LMArena Chinese | 1442 | 1526 |
| LMArena French | 1417 | 1490 |
| LMArena German | 1404 | 1492 |
| LMArena Japanese | 1391 | 1486 |
| LMArena Korean | 1360 | 1451 |
| LMArena Russian | 1423 | 1493 |
| LMArena Spanish | 1411 | 1480 |

## Instruction Following

- DeepSeek-R1: 72.0 (#143)
- Gemini 3.5 Flash: 77.0 (#30)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| LMArena Instruction Following | 1382 | 1467 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |

## Long Context

- DeepSeek-R1: 45.4 (#36)
- Gemini 3.5 Flash: 45.4 (#38)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| LMArena Longer Query | 1391 | 1482 |
| Fiction.LiveBench | 75% | — |

## Writing & Preference

- DeepSeek-R1: 61.4 (#88)
- Gemini 3.5 Flash: 65.5 (#47)

| Benchmark | DeepSeek-R1 | Gemini 3.5 Flash |
|---|---|---|
| LMArena Text | 1428 | 1482 |
| LMArena Creative Writing | 1405 | 1470 |
| LMArena Multi-Turn | 1405 | 1481 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| EQ-Bench 4 | — | 1087 |
| LiveBench Language | 48.5% | — |

## FAQ

### Is DeepSeek-R1 better than Gemini 3.5 Flash?

Gemini 3.5 Flash is the stronger model overall, scoring 54.2 to 42.3 on the Noometry Index. DeepSeek-R1 costs 3.7× less per token, which makes it the better buy when Gemini 3.5 Flash's lead doesn't matter for your workload.

### Which is cheaper, DeepSeek-R1 or Gemini 3.5 Flash?

DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Gemini 3.5 Flash lists at $1.50 and $9.

### Is DeepSeek-R1 or Gemini 3.5 Flash better for coding?

Gemini 3.5 Flash scores higher on coding benchmarks: 49.4 versus 46.3 in the Noometry coding category.

### Which has the bigger context window?

Gemini 3.5 Flash does, with 1.05M tokens against 164K.

### How many benchmarks do DeepSeek-R1 and Gemini 3.5 Flash share?

28 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Gemini 3.5 Flash has 54.