# Deepseek Coder v2 vs GPT-5.4 mini

> GPT-5.4 mini is the stronger model overall, scoring 45.0 to 35.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-coder-v2-vs-gpt-5-4-mini
- Last updated: 2026-10-10
- Shared benchmarks: 17

## Summary

- They share 17 benchmarks with published results for both. Deepseek Coder v2 scores higher in 0 categories and GPT-5.4 mini in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-5.4 mini leads 64.0 to 38.2.
- Deepseek Coder v2 has downloadable open weights; the other is API-only.

## Snapshot

| | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 35.9 | 45.0 |
| Rank | 220 | 76 |
| Context | — | 400K |
| Input $/M | — | $0.75 |
| Output $/M | — | $4.50 |
| Weights | Open | Proprietary |

## Coding

- Deepseek Coder v2: 38.1 (#183)
- GPT-5.4 mini: 45.2 (#72)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Coding | 1251 | 1438 |
| FrontierCode | — | 27% |
| LMArena WebDev | — | 1397 |
| SciCode | — | 49.9% |
| WeirdML | — | 60.3% |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| ALE-Bench | — | 1,189 |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |

## Agentic & Tool Use

- Deepseek Coder v2: —
- GPT-5.4 mini: 29.9 (#81)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| DeepResearch Bench | — | 36.3% |

## Reasoning

- Deepseek Coder v2: 23.6 (#176)
- GPT-5.4 mini: 30.4 (#85)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1424 |
| ARC-AGI-2 | — | 18.9% |
| Kagi LLM Benchmark | — | 37.9% |
| NYT Connections (extended) | — | 61.8% |
| ARC-AGI-1 | — | 63.7% |
| CritPt | — | 10% |
| Chess Puzzles | — | 24% |
| Thematic Generalization | — | 61.7% |
| Mystery Game Puzzles | — | 11% |
| DTBench | — | 80% |
| LMCA | — | 40.8% |
| Epoch Capabilities Index | — | 148.84 |
| ForecastBench | — | 57 |
| WinoGrande | 83.7% | — |

## Math

- Deepseek Coder v2: 34.9 (#190)
- GPT-5.4 mini: 45.5 (#75)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Math | 1241 | 1419 |
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 9.8% |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| ProofBench | — | 21% |
| FrontierMath (Feb 2025 set) | — | 28.3% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
| GSM8K | 94.5% | — |

## Knowledge

- Deepseek Coder v2: 32.3 (#212)
- GPT-5.4 mini: 51.5 (#67)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Expert | 1181 | 1435 |
| GPQA Diamond | — | 86.9% |
| SimpleQA Verified | — | 29.4% |
| Vectara Hallucination Rate | — | 5.5% |
| ARC (AI2) Challenge | 64.3% | — |

## Multimodal

- Deepseek Coder v2: —
- GPT-5.4 mini: 39.7 (#56)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Vision | — | 1245 |

## Multilingual

- Deepseek Coder v2: 36.3 (#240)
- GPT-5.4 mini: 51.9 (#96)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Non-English | 1182 | 1405 |
| LMArena Chinese | 1201 | 1446 |
| LMArena French | 1185 | 1440 |
| LMArena German | 1164 | 1409 |
| LMArena Japanese | 1126 | 1374 |
| LMArena Korean | 1104 | 1368 |
| LMArena Russian | 1188 | 1417 |
| LMArena Spanish | 1153 | 1405 |

## Instruction Following

- Deepseek Coder v2: 61.7 (#242)
- GPT-5.4 mini: 74.1 (#102)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Instruction Following | 1180 | 1405 |

## Long Context

- Deepseek Coder v2: 37.0 (#224)
- GPT-5.4 mini: 43.0 (#112)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Longer Query | 1219 | 1407 |

## Writing & Preference

- Deepseek Coder v2: 38.2 (#253)
- GPT-5.4 mini: 64.0 (#58)

| Benchmark | Deepseek Coder v2 | GPT-5.4 mini |
|---|---|---|
| LMArena Text | 1191 | 1412 |
| LMArena Creative Writing | 1120 | 1370 |
| LMArena Multi-Turn | 1177 | 1429 |
| EQ-Bench Creative Writing | — | 1665 |

## FAQ

### Is Deepseek Coder v2 better than GPT-5.4 mini?

GPT-5.4 mini is the stronger model overall, scoring 45.0 to 35.9 on the Noometry Index.

### Is Deepseek Coder v2 or GPT-5.4 mini better for coding?

GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 38.1 in the Noometry coding category.

### How many benchmarks do Deepseek Coder v2 and GPT-5.4 mini share?

17 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and GPT-5.4 mini has 46.