# DeepSeek V4.1 Flash vs Grok 4

> DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 48.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-1-flash-vs-grok-4
- Last updated: 2026-10-10
- Shared benchmarks: 21

## Summary

- They share 21 benchmarks with published results for both. DeepSeek V4.1 Flash scores higher in 7 categories and Grok 4 in 3 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4.1 Flash leads 66.7 to 48.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.3% for DeepSeek V4.1 Flash and 84% for Grok 4.
- DeepSeek V4.1 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 52.8 | 48.1 |
| Rank | 38 | 56 |
| Context | 1M | — |
| Input $/M | $0.15 | — |
| Output $/M | $0.60 | — |
| Weights | Open | Proprietary |

## Coding

- DeepSeek V4.1 Flash: 52.9 (#32)
- Grok 4: 50.3 (#46)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Coding | 1506 | 1408 |
| Aider Polyglot | — | 79.6% |
| LMArena WebDev | 1619 | — |
| SciCode | 51.9% | — |
| WeirdML | — | 45.7% |
| ALE-Bench | 1,092 | — |

## Agentic & Tool Use

- DeepSeek V4.1 Flash: 31.2 (#69)
- Grok 4: 32.3 (#68)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| Terminal-Bench | — | 27.2% |
| APEX-Agents | 39.5% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 21.1% |
| Cybench | — | 43% |
| DeepResearch Bench | — | 47.3% |
| BALROG | — | 43.6% |
| GDP.pdf | 19.8% | — |
| LMArena Search | — | 1142 |
| METR Time Horizons | — | 66.6% |

## Reasoning

- DeepSeek V4.1 Flash: 50.2 (#36)
- Grok 4: 36.7 (#65)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Hard Prompts | 1483 | 1409 |
| Epoch Capabilities Index | 154.9 | 146.44 |
| ARC-AGI-2 | — | 16% |
| SimpleBench | — | 60.5% |
| Kagi LLM Benchmark | — | 73.6% |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | — | 66.7% |
| CritPt | 14.3% | — |
| Chess Puzzles | — | 28% |
| Mystery Game Puzzles | 43% | — |
| DTBench | 89.9% | — |
| LMCA | 47% | — |
| Surface Evolver Bench | 46.3% | — |
| ForecastBench | — | 60.9 |

## Math

- DeepSeek V4.1 Flash: 66.7 (#25)
- Grok 4: 48.4 (#64)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.3% | 84% |
| LMArena Math | 1477 | 1422 |
| FrontierMath (Tiers 1-3) | 67.4% | — |
| FrontierMath Tier 4 | 26.8% | — |
| ProofBench | 54% | — |
| Omni-MATH | — | 60.3% |
| FrontierMath (Feb 2025 set) | — | 19.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- DeepSeek V4.1 Flash: 57.9 (#38)
- Grok 4: 53.8 (#55)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| GPQA Diamond | 89.8% | 87% |
| LMArena Expert | 1506 | 1415 |
| MMLU-Pro | — | 85.1% |
| Confabulations | — | 12.4% |
| GPQA (HELM) | — | 72.7% |

## Multimodal

- DeepSeek V4.1 Flash: 39.1 (#61)
- Grok 4: 33.7 (#94)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Vision | 1277 | 1210 |
| GeoBench | — | 45% |
| Furniture Assembly | 34.2% | — |

## Multilingual

- DeepSeek V4.1 Flash: 55.0 (#35)
- Grok 4: 51.8 (#103)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Non-English | 1448 | 1403 |
| LMArena Chinese | 1497 | 1427 |
| LMArena French | 1452 | 1418 |
| LMArena German | 1484 | 1429 |
| LMArena Japanese | 1412 | 1394 |
| LMArena Korean | 1452 | 1377 |
| LMArena Russian | 1471 | 1410 |
| LMArena Spanish | 1459 | 1420 |

## Instruction Following

- DeepSeek V4.1 Flash: 77.3 (#26)
- Grok 4: 79.2 (#5)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Instruction Following | 1474 | 1387 |
| IFEval | — | 94.9% |

## Long Context

- DeepSeek V4.1 Flash: 45.2 (#47)
- Grok 4: 63.1 (#4)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Longer Query | 1475 | 1409 |
| Fiction.LiveBench | — | 94.4% |

## Writing & Preference

- DeepSeek V4.1 Flash: 65.4 (#48)
- Grok 4: 58.5 (#116)

| Benchmark | DeepSeek V4.1 Flash | Grok 4 |
|---|---|---|
| LMArena Text | 1462 | 1411 |
| LMArena Creative Writing | 1435 | 1397 |
| LMArena Multi-Turn | 1457 | 1416 |
| Short-Story Creative Writing | — | 76.9% |
| EQ-Bench Creative Writing | 1540 | — |
| WildBench | — | 79.7% |

## FAQ

### Is DeepSeek V4.1 Flash better than Grok 4?

DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 48.1 on the Noometry Index.

### Is DeepSeek V4.1 Flash or Grok 4 better for coding?

DeepSeek V4.1 Flash scores higher on coding benchmarks: 52.9 versus 50.3 in the Noometry coding category.

### How many benchmarks do DeepSeek V4.1 Flash and Grok 4 share?

21 benchmarks have published results for both models. DeepSeek V4.1 Flash has 37 scored results on Noometry and Grok 4 has 48.