# DeepSeek V4 Flash vs Gemini 2.5 Flash-Lite

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 36.4 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-gemini-2-5-flash-lite
- Last updated: 2026-10-10
- Shared benchmarks: 23

## Summary

- They share 23 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and Gemini 2.5 Flash-Lite in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 21.5.
- The biggest single-benchmark swing is DTBench: 90.9% for DeepSeek V4 Flash and 62.8% for Gemini 2.5 Flash-Lite.
- Gemini 2.5 Flash-Lite is cheaper at $0.10 / $0.40 per million input/output tokens, against $0.15 / $0.60 for DeepSeek V4 Flash.
- Gemini 2.5 Flash-Lite accepts more context: 1.05M tokens versus 1M.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| Provider | DeepSeek | Google |
| Noometry Index | 53.6 | 36.4 |
| Rank | 32 | 208 |
| Context | 1M | 1.05M |
| Input $/M | $0.15 | $0.10 |
| Output $/M | $0.60 | $0.40 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Flash: 47.6 (#56)
- Gemini 2.5 Flash-Lite: 38.0 (#167)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| WeirdML | 63% | 35.2% |
| LMArena Coding | 1457 | 1373 |
| ALE-Bench | 1,306 | 325.9 |
| FrontierCode | 18.8% | — |
| LMArena WebDev | 1581 | — |
| SciCode | 49.9% | — |

## Agentic & Tool Use

- DeepSeek V4 Flash: —
- Gemini 2.5 Flash-Lite: 27.8 (#89)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 36.9% |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#25)
- Gemini 2.5 Flash-Lite: 21.5 (#196)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 40.5% |
| LMArena Hard Prompts | 1443 | 1377 |
| DTBench | 90.9% | 62.8% |
| LMCA | 41.7% | 18.1% |
| Epoch Capabilities Index | 154.51 | 133.93 |
| ARC-AGI-2 | 61.4% | — |
| SimpleBench | 61.1% | — |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | 89% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| Mystery Game Puzzles | 34% | — |

## Math

- DeepSeek V4 Flash: 60.0 (#31)
- Gemini 2.5 Flash-Lite: 35.5 (#144)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Math | 1427 | 1373 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| OTIS Mock AIME 2024-2025 | 94.4% | — |
| ProofBench | 56% | — |
| Omni-MATH | — | 48% |

## Knowledge

- DeepSeek V4 Flash: 55.5 (#43)
- Gemini 2.5 Flash-Lite: 32.1 (#206)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Expert | 1441 | 1373 |
| GPQA Diamond | 91% | — |
| SimpleQA Verified | 33.6% | — |
| MMLU-Pro | — | 53.7% |
| Vectara Hallucination Rate | — | 3.3% |
| GPQA (HELM) | — | 30.9% |

## Multimodal

- DeepSeek V4 Flash: —
- Gemini 2.5 Flash-Lite: 28.6 (#109)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Vision | — | 1198 |
| VPCT | — | 30% |

## Multilingual

- DeepSeek V4 Flash: 53.2 (#71)
- Gemini 2.5 Flash-Lite: 49.5 (#129)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Non-English | 1420 | 1369 |
| LMArena Chinese | 1467 | 1404 |
| LMArena French | 1438 | 1388 |
| LMArena German | 1418 | 1390 |
| LMArena Japanese | 1403 | 1359 |
| LMArena Korean | 1383 | 1358 |
| LMArena Russian | 1428 | 1373 |
| LMArena Spanish | 1437 | 1396 |

## Instruction Following

- DeepSeek V4 Flash: 75.1 (#76)
- Gemini 2.5 Flash-Lite: 70.2 (#163)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Instruction Following | 1421 | 1366 |
| IFEval | — | 81% |

## Long Context

- DeepSeek V4 Flash: 43.8 (#81)
- Gemini 2.5 Flash-Lite: 33.3 (#257)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Longer Query | 1435 | 1373 |
| Fiction.LiveBench | — | 47.2% |

## Writing & Preference

- DeepSeek V4 Flash: 64.0 (#59)
- Gemini 2.5 Flash-Lite: 57.0 (#130)

| Benchmark | DeepSeek V4 Flash | Gemini 2.5 Flash-Lite |
|---|---|---|
| LMArena Text | 1432 | 1379 |
| LMArena Creative Writing | 1403 | 1366 |
| LMArena Multi-Turn | 1449 | 1367 |
| EQ-Bench Creative Writing | 1559 | — |
| WildBench | — | 81.8% |

## FAQ

### Is DeepSeek V4 Flash better than Gemini 2.5 Flash-Lite?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 36.4 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Flash or Gemini 2.5 Flash-Lite?

Gemini 2.5 Flash-Lite is cheaper. It lists at $0.10 per million input tokens and $0.40 per million output tokens; DeepSeek V4 Flash lists at $0.15 and $0.60.

### Is DeepSeek V4 Flash or Gemini 2.5 Flash-Lite better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.6 versus 38.0 in the Noometry coding category.

### Which has the bigger context window?

Gemini 2.5 Flash-Lite does, with 1.05M tokens against 1M.

### How many benchmarks do DeepSeek V4 Flash and Gemini 2.5 Flash-Lite share?

23 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and Gemini 2.5 Flash-Lite has 33.