Model comparison
Deepseek Coder v2 vs Grok 3
Grok 3 is the stronger model overall, scoring 39.9 to 35.9 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeekSummary
- They share 17 benchmarks with published results for both. Deepseek Coder v2 scores higher in 1 category and Grok 3 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Grok 3 leads 55.8 to 38.2.
- Deepseek Coder v2 has downloadable open weights; the other is API-only.
Side by side
| Deepseek Coder v2 | Grok 3 | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 35.9 | 39.9 |
| Released | 2024-06-17 | 2025-04-09 |
| Weights | Open | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 3 leads
Deepseek Coder v2: 38.1 (#183), Grok 3: 41.9 (#115)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Coding | 1251 | 1432 |
| Aider Polyglot | — | 53.3% |
| WeirdML | — | 37.2% |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Agentic & Tool Use Not comparable
Deepseek Coder v2: —, Grok 3: 30.5 (#76)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| BALROG | — | 29.5% |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), Grok 3: 13.7 (#333)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1434 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 36.1% |
| Kagi LLM Benchmark | — | 61.3% |
| ARC-AGI-1 | — | 5.5% |
| Epoch Capabilities Index | — | 138.33 |
| WinoGrande | 83.7% | — |
Math Grok 3 leads
Deepseek Coder v2: 34.9 (#190), Grok 3: 38.0 (#145)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Math | 1241 | 1391 |
| OTIS Mock AIME 2024-2025 | — | 55.6% |
| Omni-MATH | — | 46.4% |
| MATH Level 5 | — | 88.7% |
| FrontierMath (Feb 2025 set) | — | 3.8% |
| FrontierMath Tier 4 (v1) | — | 0% |
| GSM8K | 94.5% | — |
Knowledge Grok 3 leads
Deepseek Coder v2: 32.3 (#212), Grok 3: 46.2 (#82)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Expert | 1181 | 1421 |
| GPQA Diamond | — | 75.8% |
| MMLU-Pro | — | 78.8% |
| Confabulations | — | 14.2% |
| Vectara Hallucination Rate | — | 5.8% |
| GPQA (HELM) | — | 65% |
| ARC (AI2) Challenge | 64.3% | — |
Multilingual Grok 3 leads
Deepseek Coder v2: 36.3 (#240), Grok 3: 52.3 (#87)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Non-English | 1182 | 1410 |
| LMArena Chinese | 1201 | 1448 |
| LMArena French | 1185 | 1460 |
| LMArena German | 1164 | 1431 |
| LMArena Japanese | 1126 | 1387 |
| LMArena Korean | 1104 | 1373 |
| LMArena Russian | 1188 | 1416 |
| LMArena Spanish | 1153 | 1417 |
Instruction Following Grok 3 leads
Deepseek Coder v2: 61.7 (#242), Grok 3: 75.0 (#73)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Instruction Following | 1180 | 1409 |
| IFEval | — | 88.4% |
Long Context Grok 3 leads
Deepseek Coder v2: 37.0 (#224), Grok 3: 38.7 (#192)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Longer Query | 1219 | 1439 |
| Fiction.LiveBench | — | 58.3% |
Writing & Preference Grok 3 leads
Deepseek Coder v2: 38.2 (#253), Grok 3: 55.8 (#141)
| Benchmark | Deepseek Coder v2 | Grok 3 |
|---|---|---|
| LMArena Text | 1191 | 1426 |
| LMArena Creative Writing | 1120 | 1414 |
| LMArena Multi-Turn | 1177 | 1425 |
| Short-Story Creative Writing | — | 76.4% |
| EQ-Bench Creative Writing | — | 1186 |
| WildBench | — | 84.9% |
Frequently asked questions
Is Deepseek Coder v2 better than Grok 3?
Grok 3 is the stronger model overall, scoring 39.9 to 35.9 on the Noometry Index.
Is Deepseek Coder v2 or Grok 3 better for coding?
Grok 3 scores higher on coding benchmarks: 41.9 versus 38.1 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Grok 3 share?
17 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Grok 3 has 40.