# DeepSeek V4 Flash vs Grok 3

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 39.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-grok-3
- Last updated: 2026-10-11
- Shared benchmarks: 26

## Summary

- They share 26 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 7 categories and Grok 3 in 1 category; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 13.7.
- The biggest single-benchmark swing is ARC-AGI-1: 89% for DeepSeek V4 Flash and 5.5% for Grok 3.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 53.6 | 39.9 |
| Rank | 35 | 157 |
| Context | 1M | — |
| Input $/M | $0.15 | — |
| Output $/M | $0.60 | — |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4 Flash: 47.9 (#59)
- Grok 3: 41.9 (#115)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| WeirdML | 63% | 37.2% |
| LMArena Coding | 1457 | 1432 |
| FrontierCode | 18.8% | — |
| Aider Polyglot | — | 53.3% |
| LMArena WebDev | 1582 | — |
| SciCode | 49.9% | — |
| ALE-Bench | 1,306 | — |

## Agentic & Tool Use

- DeepSeek V4 Flash: —
- Grok 3: 30.5 (#76)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| BALROG | — | 29.5% |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#30)
- Grok 3: 13.7 (#333)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| ARC-AGI-2 | 61.4% | 0% |
| SimpleBench | 61.1% | 36.1% |
| Kagi LLM Benchmark | 52.2% | 61.3% |
| ARC-AGI-1 | 89% | 5.5% |
| LMArena Hard Prompts | 1444 | 1434 |
| Epoch Capabilities Index | 154.49 | 138.33 |
| NYT Connections (extended) | 89.6% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LMCA | 41.7% | — |

## Math

- DeepSeek V4 Flash: 60.3 (#37)
- Grok 3: 38.0 (#145)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 94.4% | 55.6% |
| LMArena Math | 1427 | 1391 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| ProofBench | 56% | — |
| Omni-MATH | — | 46.4% |
| MATH Level 5 | — | 88.7% |
| FrontierMath (Feb 2025 set) | — | 3.8% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- DeepSeek V4 Flash: 55.4 (#48)
- Grok 3: 46.2 (#82)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| GPQA Diamond | 91% | 75.8% |
| LMArena Expert | 1441 | 1421 |
| SimpleQA Verified | 33.6% | — |
| MMLU-Pro | — | 78.8% |
| Confabulations | — | 14.2% |
| Vectara Hallucination Rate | — | 5.8% |
| GPQA (HELM) | — | 65% |

## Multilingual

- DeepSeek V4 Flash: 53.0 (#72)
- Grok 3: 52.3 (#87)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| LMArena Non-English | 1420 | 1410 |
| LMArena Chinese | 1468 | 1448 |
| LMArena French | 1439 | 1460 |
| LMArena German | 1418 | 1431 |
| LMArena Japanese | 1406 | 1387 |
| LMArena Korean | 1384 | 1373 |
| LMArena Russian | 1428 | 1416 |
| LMArena Spanish | 1436 | 1417 |

## Instruction Following

- DeepSeek V4 Flash: 74.9 (#81)
- Grok 3: 75.0 (#73)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| LMArena Instruction Following | 1421 | 1409 |
| IFEval | — | 88.4% |

## Long Context

- DeepSeek V4 Flash: 43.8 (#85)
- Grok 3: 38.7 (#192)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| LMArena Longer Query | 1434 | 1439 |
| Fiction.LiveBench | — | 58.3% |

## Writing & Preference

- DeepSeek V4 Flash: 63.8 (#61)
- Grok 3: 55.8 (#141)

| Benchmark | DeepSeek V4 Flash | Grok 3 |
|---|---|---|
| LMArena Text | 1432 | 1426 |
| LMArena Creative Writing | 1403 | 1414 |
| EQ-Bench Creative Writing | 1559 | 1186 |
| LMArena Multi-Turn | 1449 | 1425 |
| Short-Story Creative Writing | — | 76.4% |
| WildBench | — | 84.9% |

## FAQ

### Is DeepSeek V4 Flash better than Grok 3?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 39.9 on the Noometry Index.

### Is DeepSeek V4 Flash or Grok 3 better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 41.9 in the Noometry coding category.

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

26 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and Grok 3 has 40.