# Claude 3 Sonnet vs DeepSeek V4 Flash

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 29.0 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-sonnet-vs-deepseek-v4-flash
- Last updated: 2026-10-11
- Shared benchmarks: 22

## Summary

- They share 22 benchmarks with published results for both. Claude 3 Sonnet scores higher in 0 categories and DeepSeek V4 Flash in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Flash leads 60.3 to 10.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.5% for Claude 3 Sonnet and 94.4% for DeepSeek V4 Flash.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 29.0 | 53.6 |
| Rank | 319 | 35 |
| Context | — | 1M |
| Input $/M | — | $0.15 |
| Output $/M | — | $0.60 |
| Weights | Proprietary | Open |

## Coding

- Claude 3 Sonnet: 29.6 (#302)
- DeepSeek V4 Flash: 47.9 (#59)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| WeirdML | 10.2% | 63% |
| LMArena Coding | 1223 | 1457 |
| FrontierCode | — | 18.8% |
| LMArena WebDev | — | 1582 |
| SciCode | — | 49.9% |
| BigCodeBench Instruct | 42.7% | — |
| BigCodeBench Complete | 53.8% | — |
| ALE-Bench | — | 1,306 |
| HumanEval+ | 64% | — |
| MBPP+ | 69.3% | — |

## Reasoning

- Claude 3 Sonnet: 20.5 (#237)
- DeepSeek V4 Flash: 53.7 (#30)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| LMArena Hard Prompts | 1197 | 1444 |
| DTBench | 53.6% | 90.9% |
| Epoch Capabilities Index | 120.7 | 154.49 |
| ARC-AGI-2 | — | 61.4% |
| SimpleBench | — | 61.1% |
| Kagi LLM Benchmark | — | 52.2% |
| NYT Connections (extended) | — | 89.6% |
| ARC-AGI-1 | — | 89% |
| CritPt | — | 16.6% |
| Chess Puzzles | — | 33% |
| Mystery Game Puzzles | — | 34% |
| LMCA | — | 41.7% |
| WinoGrande | 75.1% | — |

## Math

- Claude 3 Sonnet: 10.7 (#310)
- DeepSeek V4 Flash: 60.3 (#37)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.5% | 94.4% |
| LMArena Math | 1213 | 1427 |
| FrontierMath (Tiers 1-3) | — | 57.5% |
| FrontierMath Tier 4 | — | 24.4% |
| MathArena Final-Answer Competitions | — | 76.5% |
| ProofBench | — | 56% |
| MATH Level 5 | 18.2% | — |

## Knowledge

- Claude 3 Sonnet: 21.1 (#276)
- DeepSeek V4 Flash: 55.4 (#48)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| GPQA Diamond | 40.6% | 91% |
| LMArena Expert | 1173 | 1441 |
| SimpleQA Verified | — | 33.6% |
| MMLU | 75.9% | — |

## Multimodal

- Claude 3 Sonnet: 25.2 (#125)
- DeepSeek V4 Flash: —

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| LMArena Vision | 984 | — |

## Multilingual

- Claude 3 Sonnet: 37.8 (#234)
- DeepSeek V4 Flash: 53.0 (#72)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| LMArena Non-English | 1205 | 1420 |
| LMArena Chinese | 1189 | 1468 |
| LMArena French | 1229 | 1439 |
| LMArena German | 1204 | 1418 |
| LMArena Japanese | 1131 | 1406 |
| LMArena Korean | 1128 | 1384 |
| LMArena Russian | 1227 | 1428 |
| LMArena Spanish | 1204 | 1436 |

## Instruction Following

- Claude 3 Sonnet: 62.8 (#235)
- DeepSeek V4 Flash: 74.9 (#81)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| LMArena Instruction Following | 1199 | 1421 |

## Long Context

- Claude 3 Sonnet: 36.7 (#228)
- DeepSeek V4 Flash: 43.8 (#85)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| LMArena Longer Query | 1211 | 1434 |

## Writing & Preference

- Claude 3 Sonnet: 42.1 (#238)
- DeepSeek V4 Flash: 63.8 (#61)

| Benchmark | Claude 3 Sonnet | DeepSeek V4 Flash |
|---|---|---|
| LMArena Text | 1218 | 1432 |
| LMArena Creative Writing | 1186 | 1403 |
| LMArena Multi-Turn | 1227 | 1449 |
| EQ-Bench Creative Writing | — | 1559 |

## FAQ

### Is Claude 3 Sonnet better than DeepSeek V4 Flash?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 29.0 on the Noometry Index.

### Is Claude 3 Sonnet or DeepSeek V4 Flash better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 29.6 in the Noometry coding category.

### How many benchmarks do Claude 3 Sonnet and DeepSeek V4 Flash share?

22 benchmarks have published results for both models. Claude 3 Sonnet has 30 scored results on Noometry and DeepSeek V4 Flash has 41.