# DeepSeek V4.1 Flash vs GPT-4

> DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 29.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-1-flash-vs-gpt-4
- Last updated: 2026-10-10
- Shared benchmarks: 24

## Summary

- They share 24 benchmarks with published results for both. DeepSeek V4.1 Flash scores higher in 8 categories and GPT-4 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4.1 Flash leads 66.7 to 10.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.3% for DeepSeek V4.1 Flash and 1.1% for GPT-4.
- DeepSeek V4.1 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $30 / $60 for GPT-4.
- DeepSeek V4.1 Flash accepts more context: 1M tokens versus 8K.
- DeepSeek V4.1 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 52.8 | 29.1 |
| Rank | 38 | 316 |
| Context | 1M | 8K |
| Input $/M | $0.15 | $30 |
| Output $/M | $0.60 | $60 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4.1 Flash: 52.9 (#32)
- GPT-4: 31.6 (#283)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Coding | 1506 | 1254 |
| LMArena WebDev | 1619 | — |
| SciCode | 51.9% | — |
| WeirdML | — | 12.4% |
| BigCodeBench Instruct | — | 46% |
| BigCodeBench Complete | — | 57.2% |
| ALE-Bench | 1,092 | — |
| HumanEval+ | — | 79.3% |

## Agentic & Tool Use

- DeepSeek V4.1 Flash: 31.2 (#69)
- GPT-4: —

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| APEX-Agents | 39.5% | — |
| GDP.pdf | 19.8% | — |
| METR Time Horizons | — | 36.1% |

## Reasoning

- DeepSeek V4.1 Flash: 50.2 (#36)
- GPT-4: 17.8 (#289)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Hard Prompts | 1483 | 1241 |
| Mystery Game Puzzles | 43% | 12% |
| DTBench | 89.9% | 62.7% |
| LMCA | 47% | 17.1% |
| Epoch Capabilities Index | 154.9 | 125.89 |
| NYT Connections (extended) | 89.6% | — |
| CritPt | 14.3% | — |
| Chess Puzzles | — | 4% |
| Surface Evolver Bench | 46.3% | — |
| BIG-Bench Hard | — | 75.1% |
| ForecastBench | — | 57.8 |
| HellaSwag | — | 95.3% |
| WinoGrande | — | 87.5% |

## Math

- DeepSeek V4.1 Flash: 66.7 (#25)
- GPT-4: 10.8 (#309)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.3% | 1.1% |
| LMArena Math | 1477 | 1269 |
| FrontierMath (Tiers 1-3) | 67.4% | — |
| FrontierMath Tier 4 | 26.8% | — |
| ProofBench | 54% | — |
| MATH Level 5 | — | 23% |
| GSM8K | — | 92% |

## Knowledge

- DeepSeek V4.1 Flash: 57.9 (#38)
- GPT-4: 18.4 (#282)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| GPQA Diamond | 89.8% | 35.7% |
| LMArena Expert | 1506 | 1211 |
| MMLU | — | 86.4% |
| TriviaQA | — | 84.8% |

## Multimodal

- DeepSeek V4.1 Flash: 39.1 (#61)
- GPT-4: —

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Vision | 1277 | — |
| Furniture Assembly | 34.2% | — |

## Multilingual

- DeepSeek V4.1 Flash: 55.0 (#35)
- GPT-4: 40.6 (#215)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Non-English | 1448 | 1246 |
| LMArena Chinese | 1497 | 1242 |
| LMArena French | 1452 | 1283 |
| LMArena German | 1484 | 1251 |
| LMArena Japanese | 1412 | 1209 |
| LMArena Korean | 1452 | 1184 |
| LMArena Russian | 1471 | 1251 |
| LMArena Spanish | 1459 | 1261 |

## Instruction Following

- DeepSeek V4.1 Flash: 77.3 (#26)
- GPT-4: 65.3 (#222)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Instruction Following | 1474 | 1241 |

## Long Context

- DeepSeek V4.1 Flash: 45.2 (#47)
- GPT-4: 37.7 (#212)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Longer Query | 1475 | 1244 |

## Writing & Preference

- DeepSeek V4.1 Flash: 65.4 (#48)
- GPT-4: 34.9 (#268)

| Benchmark | DeepSeek V4.1 Flash | GPT-4 |
|---|---|---|
| LMArena Text | 1462 | 1263 |
| LMArena Creative Writing | 1435 | 1244 |
| EQ-Bench Creative Writing | 1540 | 752 |
| LMArena Multi-Turn | 1457 | 1257 |

## FAQ

### Is DeepSeek V4.1 Flash better than GPT-4?

DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 29.1 on the Noometry Index.

### Which is cheaper, DeepSeek V4.1 Flash or GPT-4?

DeepSeek V4.1 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GPT-4 lists at $30 and $60.

### Is DeepSeek V4.1 Flash or GPT-4 better for coding?

DeepSeek V4.1 Flash scores higher on coding benchmarks: 52.9 versus 31.6 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4.1 Flash does, with 1M tokens against 8K.

### How many benchmarks do DeepSeek V4.1 Flash and GPT-4 share?

24 benchmarks have published results for both models. DeepSeek V4.1 Flash has 37 scored results on Noometry and GPT-4 has 38.