# DeepSeek V4.1 Flash vs o3

> DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 47.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-1-flash-vs-o3
- Last updated: 2026-10-10
- Shared benchmarks: 28

## Summary

- They share 28 benchmarks with published results for both. DeepSeek V4.1 Flash scores higher in 7 categories and o3 in 3 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4.1 Flash leads 50.2 to 32.0.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 67.4% for DeepSeek V4.1 Flash and 33.3% for o3.
- DeepSeek V4.1 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $2 / $8 for o3.
- DeepSeek V4.1 Flash accepts more context: 1M tokens versus 200K.
- DeepSeek V4.1 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 52.8 | 47.5 |
| Rank | 38 | 61 |
| Context | 1M | 200K |
| Input $/M | $0.15 | $2 |
| Output $/M | $0.60 | $8 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4.1 Flash: 52.9 (#32)
- o3: 46.8 (#64)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| LMArena Coding | 1506 | 1408 |
| ALE-Bench | 1,092 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1619 | — |
| SciCode | 51.9% | — |
| GSO | — | 8.8% |
| WeirdML | — | 52.4% |
| CadEval | — | 74% |

## Agentic & Tool Use

- DeepSeek V4.1 Flash: 31.2 (#69)
- o3: 34.5 (#44)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| APEX-Agents | 39.5% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| GDP.pdf | 19.8% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |

## Reasoning

- DeepSeek V4.1 Flash: 50.2 (#36)
- o3: 32.0 (#78)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| CritPt | 14.3% | 1.4% |
| LMArena Hard Prompts | 1483 | 1402 |
| Mystery Game Puzzles | 43% | 29% |
| DTBench | 89.9% | 84.8% |
| LMCA | 47% | 39.7% |
| Epoch Capabilities Index | 154.9 | 146.86 |
| ARC-AGI-2 | — | 6.5% |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | — | 60.8% |
| Chess Puzzles | — | 38% |
| EnigmaEval | — | 13.1% |
| Surface Evolver Bench | 46.3% | — |
| ForecastBench | — | 62.5 |

## Math

- DeepSeek V4.1 Flash: 66.7 (#25)
- o3: 50.2 (#58)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 67.4% | 33.3% |
| OTIS Mock AIME 2024-2025 | 98.3% | 84.4% |
| LMArena Math | 1477 | 1426 |
| FrontierMath Tier 4 | 26.8% | — |
| ProofBench | 54% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- DeepSeek V4.1 Flash: 57.9 (#38)
- o3: 54.6 (#52)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| GPQA Diamond | 89.8% | 81.8% |
| LMArena Expert | 1506 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| SimpleQA Verified | — | 49.4% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |

## Multimodal

- DeepSeek V4.1 Flash: 39.1 (#61)
- o3: 41.4 (#36)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| LMArena Vision | 1277 | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Furniture Assembly | 34.2% | — |

## Multilingual

- DeepSeek V4.1 Flash: 55.0 (#35)
- o3: 51.7 (#105)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| LMArena Non-English | 1448 | 1401 |
| LMArena Chinese | 1497 | 1437 |
| LMArena French | 1452 | 1430 |
| LMArena German | 1484 | 1420 |
| LMArena Japanese | 1412 | 1403 |
| LMArena Korean | 1452 | 1370 |
| LMArena Russian | 1471 | 1406 |
| LMArena Spanish | 1459 | 1395 |

## Instruction Following

- DeepSeek V4.1 Flash: 77.3 (#26)
- o3: 72.8 (#127)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| LMArena Instruction Following | 1474 | 1368 |
| IFEval | — | 86.9% |

## Long Context

- DeepSeek V4.1 Flash: 45.2 (#47)
- o3: 53.3 (#6)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| LMArena Longer Query | 1475 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |

## Writing & Preference

- DeepSeek V4.1 Flash: 65.4 (#48)
- o3: 63.5 (#64)

| Benchmark | DeepSeek V4.1 Flash | o3 |
|---|---|---|
| LMArena Text | 1462 | 1410 |
| LMArena Creative Writing | 1435 | 1359 |
| EQ-Bench Creative Writing | 1540 | 1676 |
| LMArena Multi-Turn | 1457 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |

## FAQ

### Is DeepSeek V4.1 Flash better than o3?

DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 47.5 on the Noometry Index.

### Which is cheaper, DeepSeek V4.1 Flash or o3?

DeepSeek V4.1 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; o3 lists at $2 and $8.

### Is DeepSeek V4.1 Flash or o3 better for coding?

DeepSeek V4.1 Flash scores higher on coding benchmarks: 52.9 versus 46.8 in the Noometry coding category.

### Which has the bigger context window?

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

### How many benchmarks do DeepSeek V4.1 Flash and o3 share?

28 benchmarks have published results for both models. DeepSeek V4.1 Flash has 37 scored results on Noometry and o3 has 63.