# DeepSeek V4 Flash vs o3-mini

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 36.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-o3-mini
- Last updated: 2026-10-10
- Shared benchmarks: 33

## Summary

- They share 33 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 7 categories and o3-mini in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 16.3.
- The biggest single-benchmark swing is ARC-AGI-2: 61.4% for DeepSeek V4 Flash and 3% for o3-mini.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $1.10 / $4.40 for o3-mini.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 200K.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 53.6 | 36.7 |
| Rank | 35 | 212 |
| Context | 1M | 200K |
| Input $/M | $0.15 | $1.10 |
| Output $/M | $0.60 | $4.40 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Flash: 47.9 (#59)
- o3-mini: 40.8 (#132)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| SciCode | 49.9% | 39.8% |
| WeirdML | 63% | 43.7% |
| LMArena Coding | 1457 | 1378 |
| FrontierCode | 18.8% | — |
| Aider Polyglot | — | 60.4% |
| LMArena WebDev | 1582 | — |
| GSO | — | 1.3% |
| LiveBench Coding | — | 82.7% |
| CadEval | — | 54% |
| ALE-Bench | 1,306 | — |

## Agentic & Tool Use

- DeepSeek V4 Flash: —
- o3-mini: 29.6 (#84)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| Cybench | — | 22.5% |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#30)
- o3-mini: 16.3 (#305)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| ARC-AGI-2 | 61.4% | 3% |
| SimpleBench | 61.1% | 22.8% |
| ARC-AGI-1 | 89% | 34.5% |
| CritPt | 16.6% | 0.3% |
| Chess Puzzles | 33% | 17% |
| LMArena Hard Prompts | 1444 | 1366 |
| Mystery Game Puzzles | 34% | 7% |
| DTBench | 90.9% | 68.8% |
| LMCA | 41.7% | 19% |
| Epoch Capabilities Index | 154.49 | 140.34 |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 89.6% | — |
| LiveBench Reasoning | — | 89.6% |
| LiveBench Data Analysis | — | 70.6% |
| ForecastBench | — | 59.6 |
| LiveBench | — | 75.9% |

## Math

- DeepSeek V4 Flash: 60.3 (#37)
- o3-mini: 28.1 (#244)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.5% | 18.6% |
| FrontierMath Tier 4 | 24.4% | 0% |
| OTIS Mock AIME 2024-2025 | 94.4% | 76.9% |
| LMArena Math | 1427 | 1396 |
| MathArena Final-Answer Competitions | 76.5% | — |
| ProofBench | 56% | — |
| LiveBench Math | — | 77.3% |
| MATH Level 5 | — | 96.5% |
| FrontierMath (Feb 2025 set) | — | 12.4% |
| FrontierMath Tier 4 (v1) | — | 4.2% |

## Knowledge

- DeepSeek V4 Flash: 55.4 (#48)
- o3-mini: 38.3 (#146)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| GPQA Diamond | 91% | 77% |
| SimpleQA Verified | 33.6% | 15.3% |
| LMArena Expert | 1441 | 1364 |
| Confabulations | — | 17.9% |

## Multilingual

- DeepSeek V4 Flash: 53.0 (#72)
- o3-mini: 45.7 (#164)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| LMArena Non-English | 1420 | 1319 |
| LMArena Chinese | 1468 | 1379 |
| LMArena French | 1439 | 1334 |
| LMArena German | 1418 | 1303 |
| LMArena Japanese | 1406 | 1286 |
| LMArena Korean | 1384 | 1314 |
| LMArena Russian | 1428 | 1304 |
| LMArena Spanish | 1436 | 1321 |

## Instruction Following

- DeepSeek V4 Flash: 74.9 (#81)
- o3-mini: 75.1 (#72)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| LMArena Instruction Following | 1421 | 1337 |
| LiveBench Instruction Following | — | 84.4% |

## Long Context

- DeepSeek V4 Flash: 43.8 (#85)
- o3-mini: 33.8 (#256)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| LMArena Longer Query | 1434 | 1343 |
| Fiction.LiveBench | — | 50% |

## Writing & Preference

- DeepSeek V4 Flash: 63.8 (#61)
- o3-mini: 50.3 (#182)

| Benchmark | DeepSeek V4 Flash | o3-mini |
|---|---|---|
| LMArena Text | 1432 | 1337 |
| LMArena Creative Writing | 1403 | 1286 |
| LMArena Multi-Turn | 1449 | 1320 |
| Short-Story Creative Writing | — | 61.7% |
| EQ-Bench Creative Writing | 1559 | — |
| LiveBench Language | — | 50.7% |

## FAQ

### Is DeepSeek V4 Flash better than o3-mini?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 36.7 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Flash or o3-mini?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; o3-mini lists at $1.10 and $4.40.

### Is DeepSeek V4 Flash or o3-mini better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 40.8 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 200K.

### How many benchmarks do DeepSeek V4 Flash and o3-mini share?

33 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and o3-mini has 51.