# DeepSeek Coder 6.7B vs GPT-5.4

> GPT-5.4 has enough public results to be ranked (#16); DeepSeek Coder 6.7B does not yet, so treat this comparison as directional.

- Canonical page: https://noometry.com/compare/deepseek-coder-6-7b-vs-gpt-5-4
- Last updated: 2026-10-11
- Shared benchmarks: 1

## Summary

- They share 1 benchmark with published results for both. DeepSeek Coder 6.7B scores higher in 0 categories and GPT-5.4 in 1 category; one gap is clear of the uncertainty.
- The widest gap is in coding, where GPT-5.4 leads 52.6 to 35.8.
- DeepSeek Coder 6.7B has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 37.6 | 59.4 |
| Rank | — | 16 |
| Context | — | 1.05M |
| Input $/M | — | $2.50 |
| Output $/M | — | $15 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek Coder 6.7B: 35.8 (#219)
- GPT-5.4: 52.6 (#33)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| SWE-bench Verified | — | 76.9% |
| DeepSWE | — | 51.8% |
| LMArena WebDev | — | 1465 |
| SciCode | — | 56.6% |
| GSO | — | 31.4% |
| WeirdML | — | 77.7% |
| BigCodeBench Instruct | 35.5% | — |
| LMArena Coding | — | 1497 |
| MirrorCode | — | 15.6% |
| BigCodeBench Complete | 43.8% | — |
| ALE-Bench | — | 1,607 |
| AlgoTune | — | 1.85 |
| HumanEval+ | 71.3% | — |
| MBPP+ | 65.6% | — |

## Agentic & Tool Use

- DeepSeek Coder 6.7B: —
- GPT-5.4: 46.5 (#13)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| Terminal-Bench | — | 81.8% |
| APEX-Agents | — | 52.4% |
| τ²-bench Banking | — | 39.4% |
| DeepResearch Bench | — | 35.1% |
| PostTrainBench | — | 19% |
| GBAEval | — | 45.1% |
| LMArena Search | — | 1197 |
| METR Time Horizons | — | 74.3% |
| Vending-Bench 2 | — | 6,144 |

## Reasoning

- DeepSeek Coder 6.7B: —
- GPT-5.4: 61.8 (#19)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| Epoch Capabilities Index | 89.62 | 156.81 |
| ARC-AGI-2 | — | 74% |
| Kagi LLM Benchmark | — | 63.8% |
| NYT Connections (extended) | — | 91.3% |
| ARC-AGI-1 | — | 93.7% |
| CritPt | — | 23.4% |
| Chess Puzzles | — | 44% |
| EnigmaEval | — | 16% |
| Thematic Generalization | — | 80% |
| EBR-Bench | — | 25.4% |
| LMArena Hard Prompts | — | 1485 |
| Mystery Game Puzzles | — | 37% |
| DTBench | — | 94.4% |
| LMCA | — | 52% |
| ForecastBench | — | 59.5 |
| WinoGrande | 57.6% | — |

## Math

- DeepSeek Coder 6.7B: —
- GPT-5.4: 73.5 (#19)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 78.6% |
| FrontierMath Tier 4 | — | 49% |
| MathArena Final-Answer Competitions | — | 83.1% |
| OTIS Mock AIME 2024-2025 | — | 97.8% |
| ProofBench | — | 56% |
| LMArena Math | — | 1488 |
| FrontierMath (Feb 2025 set) | — | 47.6% |
| FrontierMath Tier 4 (v1) | — | 27.1% |
| GSM8K | 21.3% | — |

## Knowledge

- DeepSeek Coder 6.7B: —
- GPT-5.4: 65.3 (#14)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| GPQA Diamond | — | 93.3% |
| Humanity's Last Exam | — | 36.2% |
| SimpleQA Verified | — | 45.1% |
| Vectara Hallucination Rate | — | 7% |
| LMArena Expert | — | 1507 |
| ARC (AI2) Challenge | 36.4% | — |
| MMLU | 36.4% | — |

## Multimodal

- DeepSeek Coder 6.7B: —
- GPT-5.4: 43.7 (#20)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| LMArena Vision | — | 1303 |
| Blueprint-Bench 2 | — | 27.1% |
| Furniture Assembly | — | 37.5% |
| LMArena Document | — | 1471 |

## Multilingual

- DeepSeek Coder 6.7B: —
- GPT-5.4: 56.2 (#23)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| LMArena Non-English | — | 1465 |
| LMArena Chinese | — | 1519 |
| LMArena French | — | 1493 |
| LMArena German | — | 1472 |
| LMArena Japanese | — | 1485 |
| LMArena Korean | — | 1448 |
| LMArena Russian | — | 1480 |
| LMArena Spanish | — | 1454 |

## Instruction Following

- DeepSeek Coder 6.7B: —
- GPT-5.4: 77.1 (#27)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| LMArena Instruction Following | — | 1469 |

## Long Context

- DeepSeek Coder 6.7B: —
- GPT-5.4: 50.3 (#8)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| CL-bench | — | 27.9% |
| CL-bench Life | — | 21.7% |
| LMArena Longer Query | — | 1473 |

## Writing & Preference

- DeepSeek Coder 6.7B: —
- GPT-5.4: 71.9 (#17)

| Benchmark | DeepSeek Coder 6.7B | GPT-5.4 |
|---|---|---|
| LMArena Text | — | 1469 |
| LMArena Creative Writing | — | 1439 |
| EQ-Bench Creative Writing | — | 1840 |
| EQ-Bench 4 | — | 1272 |
| LMArena Multi-Turn | — | 1482 |

## FAQ

### Is DeepSeek Coder 6.7B better than GPT-5.4?

GPT-5.4 has enough public results to be ranked (#16); DeepSeek Coder 6.7B does not yet, so treat this comparison as directional.

### Is DeepSeek Coder 6.7B or GPT-5.4 better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.6 versus 35.8 in the Noometry coding category.

### How many benchmarks do DeepSeek Coder 6.7B and GPT-5.4 share?

1 benchmark has published results for both models. DeepSeek Coder 6.7B has 9 scored results on Noometry and GPT-5.4 has 68.