# DeepSeek V4.1 Flash vs GPT-5.4 mini

> DeepSeek V4.1 Flash is the stronger model overall, scoring 50.3 to 44.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-1-flash-vs-gpt-5-4-mini
- 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 7 categories and GPT-5.4 mini in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4.1 Flash leads 46.2 to 30.3.
- The biggest single-benchmark swing is ProofBench: 54% for DeepSeek V4.1 Flash and 21% for GPT-5.4 mini.
- DeepSeek V4.1 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
- DeepSeek V4.1 Flash accepts more context: 1M tokens versus 400K.
- DeepSeek V4.1 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 50.3 | 44.8 |
| Rank | 40 | 69 |
| Context | 1M | 400K |
| Input $/M | $0.15 | $0.75 |
| Output $/M | $0.60 | $4.50 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4.1 Flash: 52.8 (#29)
- GPT-5.4 mini: 45.0 (#67)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena WebDev | 1620 | 1397 |
| SciCode | 51.9% | 49.9% |
| LMArena Coding | 1511 | 1438 |
| ALE-Bench | 1,092 | 1,189 |
| FrontierCode | — | 27% |
| WeirdML | — | 60.3% |

## Agentic & Tool Use

- DeepSeek V4.1 Flash: —
- GPT-5.4 mini: 29.6 (#74)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| DeepResearch Bench | — | 36.3% |

## Reasoning

- DeepSeek V4.1 Flash: 46.2 (#39)
- GPT-5.4 mini: 30.3 (#77)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| NYT Connections (extended) | 89.6% | 61.8% |
| CritPt | 14.3% | 10% |
| LMArena Hard Prompts | 1485 | 1424 |
| DTBench | 89.9% | 80% |
| LMCA | 47% | 40.8% |
| Epoch Capabilities Index | 155.01 | 148.83 |
| ARC-AGI-2 | — | 18.9% |
| Kagi LLM Benchmark | — | 37.9% |
| ARC-AGI-1 | — | 63.7% |
| Chess Puzzles | — | 24% |
| Thematic Generalization | — | 61.7% |
| Mystery Game Puzzles | — | 11% |
| Surface Evolver Bench | 46.3% | — |
| ForecastBench | — | 57 |

## Math

- DeepSeek V4.1 Flash: 48.4 (#52)
- GPT-5.4 mini: 45.4 (#63)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| ProofBench | 54% | 21% |
| LMArena Math | 1493 | 1418 |
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 9.8% |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| FrontierMath (Feb 2025 set) | — | 28.3% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- DeepSeek V4.1 Flash: 41.6 (#97)
- GPT-5.4 mini: 51.4 (#62)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena Expert | 1505 | 1434 |
| GPQA Diamond | — | 86.9% |
| SimpleQA Verified | — | 29.4% |
| Vectara Hallucination Rate | — | 5.5% |

## Multimodal

- DeepSeek V4.1 Flash: —
- GPT-5.4 mini: 39.0 (#51)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena Vision | — | 1245 |

## Multilingual

- DeepSeek V4.1 Flash: 55.5 (#29)
- GPT-5.4 mini: 52.1 (#94)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena Non-English | 1452 | 1405 |
| LMArena Chinese | 1506 | 1446 |
| LMArena French | 1448 | 1440 |
| LMArena Russian | 1473 | 1417 |
| LMArena Spanish | 1499 | 1403 |
| LMArena German | — | 1407 |
| LMArena Japanese | — | 1375 |
| LMArena Korean | — | 1369 |

## Instruction Following

- DeepSeek V4.1 Flash: 77.6 (#24)
- GPT-5.4 mini: 74.3 (#99)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena Instruction Following | 1475 | 1405 |

## Long Context

- DeepSeek V4.1 Flash: 45.2 (#46)
- GPT-5.4 mini: 42.9 (#107)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena Longer Query | 1476 | 1407 |

## Writing & Preference

- DeepSeek V4.1 Flash: 65.8 (#43)
- GPT-5.4 mini: 64.2 (#56)

| Benchmark | DeepSeek V4.1 Flash | GPT-5.4 mini |
|---|---|---|
| LMArena Text | 1465 | 1411 |
| LMArena Creative Writing | 1436 | 1370 |
| EQ-Bench Creative Writing | 1540 | 1665 |
| LMArena Multi-Turn | 1460 | 1429 |

## FAQ

### Is DeepSeek V4.1 Flash better than GPT-5.4 mini?

DeepSeek V4.1 Flash is the stronger model overall, scoring 50.3 to 44.8 on the Noometry Index.

### Which is cheaper, DeepSeek V4.1 Flash or GPT-5.4 mini?

DeepSeek V4.1 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.

### Is DeepSeek V4.1 Flash or GPT-5.4 mini better for coding?

DeepSeek V4.1 Flash scores higher on coding benchmarks: 52.8 versus 45.0 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4.1 Flash does, with 1M tokens against 400K.

### How many benchmarks do DeepSeek V4.1 Flash and GPT-5.4 mini share?

24 benchmarks have published results for both models. DeepSeek V4.1 Flash has 25 scored results on Noometry and GPT-5.4 mini has 46.