# DeepSeek-V2 (MoE-236B, May 2024) vs Grok-2 (Dec 2024)

> Grok-2 (Dec 2024) has enough public results to be ranked (#233); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.

- Canonical page: https://noometry.com/compare/deepseek-v2-vs-grok-2
- Last updated: 2026-10-10
- Shared benchmarks: 1

## Summary

- They share 1 benchmark with published results for both. DeepSeek-V2 (MoE-236B, May 2024) scores higher in 1 category and Grok-2 (Dec 2024) in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V2 (MoE-236B, May 2024) leads 39.8 to 32.9.
- DeepSeek-V2 (MoE-236B, May 2024) has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 39.7 | 33.4 |
| Rank | — | 233 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Open | Proprietary |

## Coding

- DeepSeek-V2 (MoE-236B, May 2024): 39.8 (#141)
- Grok-2 (Dec 2024): 32.9 (#254)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| WeirdML | — | 22.2% |
| BigCodeBench Instruct | 48.9% | — |
| LiveBench Coding | — | 46.4% |
| LMArena Coding | — | 1288 |
| BigCodeBench Complete | 59.4% | — |

## Reasoning

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 16.4 (#294)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| Epoch Capabilities Index | 124.78 | 130.48 |
| SimpleBench | — | 22.7% |
| LiveBench Reasoning | — | 54.8% |
| LMArena Hard Prompts | — | 1272 |
| DTBench | — | 65.2% |
| LiveBench Data Analysis | — | 54.5% |
| BIG-Bench Hard | 78.8% | — |
| HellaSwag | 87.1% | — |
| LiveBench | — | 54.3% |
| PIQA | 83.9% | — |
| WinoGrande | 86.3% | — |

## Math

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 19.6 (#278)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 11.5% |
| LiveBench Math | — | 54.9% |
| LMArena Math | — | 1283 |
| MATH Level 5 | — | 63.5% |
| FrontierMath (Feb 2025 set) | — | 0.7% |

## Knowledge

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 29.8 (#225)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| GPQA Diamond | — | 53.8% |
| Confabulations | — | 20.1% |
| LMArena Expert | — | 1254 |
| ARC (AI2) Challenge | 92.2% | — |
| MMLU | 78.4% | — |
| TriviaQA | 80% | — |

## Multilingual

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 43.3 (#183)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Non-English | — | 1282 |
| LMArena Chinese | — | 1289 |
| LMArena French | — | 1317 |
| LMArena German | — | 1287 |
| LMArena Japanese | — | 1244 |
| LMArena Korean | — | 1237 |
| LMArena Russian | — | 1286 |
| LMArena Spanish | — | 1281 |

## Instruction Following

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 67.1 (#197)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| LiveBench Instruction Following | — | 69.6% |
| LMArena Instruction Following | — | 1270 |

## Long Context

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 38.7 (#186)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Longer Query | — | 1276 |

## Writing & Preference

- DeepSeek-V2 (MoE-236B, May 2024): —
- Grok-2 (Dec 2024): 48.8 (#192)

| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Text | — | 1305 |
| LMArena Creative Writing | — | 1284 |
| Short-Story Creative Writing | — | 63.6% |
| LMArena Multi-Turn | — | 1290 |
| LiveBench Language | — | 45.6% |

## FAQ

### Is DeepSeek-V2 (MoE-236B, May 2024) better than Grok-2 (Dec 2024)?

Grok-2 (Dec 2024) has enough public results to be ranked (#233); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.

### Is DeepSeek-V2 (MoE-236B, May 2024) or Grok-2 (Dec 2024) better for coding?

DeepSeek-V2 (MoE-236B, May 2024) scores higher on coding benchmarks: 39.8 versus 32.9 in the Noometry coding category.

### How many benchmarks do DeepSeek-V2 (MoE-236B, May 2024) and Grok-2 (Dec 2024) share?

1 benchmark has published results for both models. DeepSeek-V2 (MoE-236B, May 2024) has 10 scored results on Noometry and Grok-2 (Dec 2024) has 34.