# DeepSeek V4 Pro vs GPT-5.4 nano

> DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 41.9 on the Noometry Index. GPT-5.4 nano costs 2.1× less per token, which makes it the better buy when DeepSeek V4 Pro's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/deepseek-v4-pro-vs-gpt-5-4-nano
- Last updated: 2026-10-12
- Shared benchmarks: 37

## Summary

- They share 37 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and GPT-5.4 nano in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 23.7.
- The biggest single-benchmark swing is ARC-AGI-2: 61.3% for DeepSeek V4 Pro and 5.7% for GPT-5.4 nano.
- GPT-5.4 nano is cheaper at $0.20 / $1.25 per million input/output tokens, against $0.66 / $1.98 for DeepSeek V4 Pro.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 400K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 54.3 | 41.9 |
| Rank | 31 | 125 |
| Context | 1M | 400K |
| Input $/M | $0.66 | $0.20 |
| Output $/M | $1.98 | $1.25 |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Pro: 52.4 (#34)
- GPT-5.4 nano: 43.6 (#84)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| SciCode | 51% | 46.9% |
| WeirdML | 66.2% | 49.2% |
| LMArena Coding | 1470 | 1405 |
| ALE-Bench | 1,403 | 1,005 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| LMArena WebDev | 1582 | — |

## Agentic & Tool Use

- DeepSeek V4 Pro: 32.8 (#58)
- GPT-5.4 nano: —

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |

## Reasoning

- DeepSeek V4 Pro: 56.5 (#24)
- GPT-5.4 nano: 23.7 (#173)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| ARC-AGI-2 | 61.3% | 5.7% |
| Kagi LLM Benchmark | 53.5% | 39.7% |
| ARC-AGI-1 | 90.5% | 51.5% |
| CritPt | 18% | 9.3% |
| Chess Puzzles | 47% | 30% |
| LMArena Hard Prompts | 1461 | 1381 |
| Mystery Game Puzzles | 43% | 9% |
| DTBench | 93.9% | 80.3% |
| LMCA | 45.5% | 36.9% |
| Epoch Capabilities Index | 155.31 | 145.81 |
| ForecastBench | 56.1 | 57.3 |
| NYT Connections (extended) | 91.3% | — |
| Surface Evolver Bench | 40% | — |

## Math

- DeepSeek V4 Pro: 64.8 (#30)
- GPT-5.4 nano: 40.9 (#88)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | 44.9% |
| FrontierMath Tier 4 | 26.8% | 12.2% |
| OTIS Mock AIME 2024-2025 | 98.6% | 87.8% |
| ProofBench | 50% | 5% |
| LMArena Math | 1455 | 1406 |
| MathArena Final-Answer Competitions | 76.6% | — |
| FrontierMath (Feb 2025 set) | — | 25.9% |
| FrontierMath Tier 4 (v1) | — | 6.3% |

## Knowledge

- DeepSeek V4 Pro: 59.5 (#31)
- GPT-5.4 nano: 41.9 (#103)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| GPQA Diamond | 91.7% | 78.5% |
| SimpleQA Verified | 52.9% | 11.7% |
| Vectara Hallucination Rate | 8.6% | 3.1% |
| LMArena Expert | 1464 | 1396 |

## Multimodal

- DeepSeek V4 Pro: —
- GPT-5.4 nano: 36.7 (#78)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| LMArena Vision | — | 1196 |

## Multilingual

- DeepSeek V4 Pro: 54.4 (#45)
- GPT-5.4 nano: 48.6 (#140)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| LMArena Non-English | 1439 | 1359 |
| LMArena Chinese | 1486 | 1392 |
| LMArena French | 1472 | 1396 |
| LMArena German | 1458 | 1367 |
| LMArena Japanese | 1445 | 1343 |
| LMArena Korean | 1447 | 1320 |
| LMArena Russian | 1453 | 1363 |
| LMArena Spanish | 1458 | 1371 |

## Instruction Following

- DeepSeek V4 Pro: 76.1 (#47)
- GPT-5.4 nano: 71.9 (#144)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| LMArena Instruction Following | 1448 | 1362 |

## Long Context

- DeepSeek V4 Pro: 45.0 (#51)
- GPT-5.4 nano: 41.6 (#137)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| LMArena Longer Query | 1458 | 1366 |
| CL-bench Life | 13.5% | — |

## Writing & Preference

- DeepSeek V4 Pro: 65.5 (#46)
- GPT-5.4 nano: 55.7 (#142)

| Benchmark | DeepSeek V4 Pro | GPT-5.4 nano |
|---|---|---|
| LMArena Text | 1451 | 1372 |
| LMArena Creative Writing | 1446 | 1314 |
| LMArena Multi-Turn | 1467 | 1382 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |

## FAQ

### Is DeepSeek V4 Pro better than GPT-5.4 nano?

DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 41.9 on the Noometry Index. GPT-5.4 nano costs 2.1× less per token, which makes it the better buy when DeepSeek V4 Pro's lead doesn't matter for your workload.

### Which is cheaper, DeepSeek V4 Pro or GPT-5.4 nano?

GPT-5.4 nano is cheaper. It lists at $0.20 per million input tokens and $1.25 per million output tokens; DeepSeek V4 Pro lists at $0.66 and $1.98.

### Is DeepSeek V4 Pro or GPT-5.4 nano better for coding?

DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 43.6 in the Noometry coding category.

### Which has the bigger context window?

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

### How many benchmarks do DeepSeek V4 Pro and GPT-5.4 nano share?

37 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and GPT-5.4 nano has 40.