# DeepSeek V4 Pro vs Llama 4 Scout

> DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 27.7 on the Noometry Index. Llama 4 Scout costs 6.6× 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-llama-4-scout
- Last updated: 2026-10-10
- Shared benchmarks: 30

## Summary

- They share 30 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 9 categories and Llama 4 Scout in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 9.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.6% for DeepSeek V4 Pro and 7.8% for Llama 4 Scout.
- Llama 4 Scout is cheaper at $0.10 / $0.30 per million input/output tokens, against $0.66 / $1.98 for DeepSeek V4 Pro.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 128K.

## Snapshot

| | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 54.3 | 27.7 |
| Rank | 31 | 330 |
| Context | 1M | 128K |
| Input $/M | $0.66 | $0.10 |
| Output $/M | $1.98 | $0.30 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Pro: 52.4 (#34)
- Llama 4 Scout: 20.2 (#339)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| SciCode | 51% | 17% |
| LMArena Coding | 1470 | 1286 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| SWE-bench Verified (bash only) | — | 9.1% |
| LMArena WebDev | 1582 | — |
| WeirdML | 66.2% | — |
| BigCodeBench Complete | — | 43.1% |
| ALE-Bench | 1,403 | — |

## Agentic & Tool Use

- DeepSeek V4 Pro: 32.8 (#58)
- Llama 4 Scout: 24.6 (#119)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Berkeley Function Calling Leaderboard | — | 28.1% |
| Vending-Bench 2 | 3,285 | — |

## Reasoning

- DeepSeek V4 Pro: 56.5 (#24)
- Llama 4 Scout: 9.1 (#345)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| ARC-AGI-2 | 61.3% | 0% |
| Kagi LLM Benchmark | 53.5% | 36.9% |
| ARC-AGI-1 | 90.5% | 0.5% |
| CritPt | 18% | 0% |
| LMArena Hard Prompts | 1461 | 1266 |
| DTBench | 93.9% | 57.9% |
| LMCA | 45.5% | 12% |
| Epoch Capabilities Index | 155.31 | 129.64 |
| ForecastBench | 56.1 | 57.5 |
| NYT Connections (extended) | 91.3% | — |
| Chess Puzzles | 47% | — |
| Mystery Game Puzzles | 43% | — |
| Surface Evolver Bench | 40% | — |

## Math

- DeepSeek V4 Pro: 64.8 (#30)
- Llama 4 Scout: 19.6 (#286)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 7.8% |
| LMArena Math | 1455 | 1287 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| ProofBench | 50% | — |
| Omni-MATH | — | 37.3% |
| MATH Level 5 | — | 62.3% |
| FrontierMath (Feb 2025 set) | — | 0% |

## Knowledge

- DeepSeek V4 Pro: 59.5 (#31)
- Llama 4 Scout: 31.9 (#217)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| GPQA Diamond | 91.7% | 51.8% |
| Vectara Hallucination Rate | 8.6% | 7.7% |
| LMArena Expert | 1464 | 1235 |
| SimpleQA Verified | 52.9% | — |
| MMLU-Pro | — | 74.2% |
| GPQA (HELM) | — | 50.7% |

## Multimodal

- DeepSeek V4 Pro: —
- Llama 4 Scout: 32.2 (#102)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| LMArena Vision | — | 1118 |
| SpatialViz-Bench | — | 34.2% |

## Multilingual

- DeepSeek V4 Pro: 54.4 (#45)
- Llama 4 Scout: 41.0 (#212)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| LMArena Non-English | 1439 | 1252 |
| LMArena Chinese | 1486 | 1255 |
| LMArena French | 1472 | 1282 |
| LMArena German | 1458 | 1272 |
| LMArena Japanese | 1445 | 1206 |
| LMArena Korean | 1447 | 1207 |
| LMArena Russian | 1453 | 1263 |
| LMArena Spanish | 1458 | 1278 |

## Instruction Following

- DeepSeek V4 Pro: 76.1 (#47)
- Llama 4 Scout: 65.8 (#217)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| LMArena Instruction Following | 1448 | 1248 |
| IFEval | — | 81.8% |

## Long Context

- DeepSeek V4 Pro: 45.0 (#51)
- Llama 4 Scout: 27.5 (#294)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| LMArena Longer Query | 1458 | 1265 |
| Fiction.LiveBench | — | 36% |
| CL-bench Life | 13.5% | — |

## Writing & Preference

- DeepSeek V4 Pro: 65.5 (#46)
- Llama 4 Scout: 37.0 (#261)

| Benchmark | DeepSeek V4 Pro | Llama 4 Scout |
|---|---|---|
| LMArena Text | 1451 | 1279 |
| LMArena Creative Writing | 1446 | 1249 |
| EQ-Bench Creative Writing | 1553 | 783 |
| LMArena Multi-Turn | 1467 | 1280 |
| WildBench | — | 78% |
| EQ-Bench 4 | 1166 | — |

## FAQ

### Is DeepSeek V4 Pro better than Llama 4 Scout?

DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 27.7 on the Noometry Index. Llama 4 Scout costs 6.6× 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 Llama 4 Scout?

Llama 4 Scout is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; DeepSeek V4 Pro lists at $0.66 and $1.98.

### Is DeepSeek V4 Pro or Llama 4 Scout better for coding?

DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 20.2 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Pro does, with 1M tokens against 128K.

### How many benchmarks do DeepSeek V4 Pro and Llama 4 Scout share?

30 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Llama 4 Scout has 43.