Model comparison
DeepSeek LLM 67B vs Grok-3 mini
Grok-3 mini is the stronger model overall, scoring 41.2 to 24.9 on the Noometry Index.
Last verified . 14 shared benchmarks.
DeepSeekSummary
- They share 14 benchmarks with published results for both. DeepSeek LLM 67B scores higher in 1 category and Grok-3 mini in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok-3 mini leads 46.4 to 7.0.
- The biggest single-benchmark swing is MATH Level 5: 6.4% for DeepSeek LLM 67B and 90.9% for Grok-3 mini.
- DeepSeek LLM 67B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek LLM 67B | Grok-3 mini | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 24.9 | 41.2 |
| Released | 2023-11-29 | 2025-04-09 |
| Weights | Open | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 15 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok-3 mini leads
DeepSeek LLM 67B: 31.9 (#278), Grok-3 mini: 40.8 (#131)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| LMArena Coding | 1096 | 1379 |
| Aider Polyglot | — | 49.3% |
| WeirdML | — | 42.6% |
Reasoning DeepSeek LLM 67B leads
DeepSeek LLM 67B: 16.5 (#304), Grok-3 mini: 13.6 (#334)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| LMArena Hard Prompts | 1070 | 1375 |
| Epoch Capabilities Index | 110.5 | 140.35 |
| ARC-AGI-2 | — | 0.4% |
| Kagi LLM Benchmark | — | 61.3% |
| ARC-AGI-1 | — | 16.5% |
| Chess Puzzles | 0% | — |
Math Grok-3 mini leads
DeepSeek LLM 67B: 8.7 (#324), Grok-3 mini: 42.1 (#85)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.8% | 77.8% |
| LMArena Math | 1108 | 1386 |
| MATH Level 5 | 6.4% | 90.9% |
| Omni-MATH | — | 31.8% |
| FrontierMath (Feb 2025 set) | — | 5.9% |
Knowledge Grok-3 mini leads
DeepSeek LLM 67B: 7.0 (#313), Grok-3 mini: 46.4 (#81)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| GPQA Diamond | 24.6% | 76.3% |
| MMLU-Pro | — | 79.9% |
| Confabulations | — | 10.8% |
| GPQA (HELM) | — | 67.5% |
| LMArena Expert | — | 1395 |
Multilingual Grok-3 mini leads
DeepSeek LLM 67B: 29.4 (#267), Grok-3 mini: 48.1 (#145)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| LMArena Non-English | 1073 | 1352 |
| LMArena Chinese | 1132 | 1387 |
| LMArena French | — | 1357 |
| LMArena German | — | 1349 |
| LMArena Japanese | — | 1342 |
| LMArena Korean | — | 1335 |
| LMArena Russian | — | 1353 |
| LMArena Spanish | — | 1381 |
Instruction Following Grok-3 mini leads
DeepSeek LLM 67B: 55.4 (#277), Grok-3 mini: 78.5 (#9)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| LMArena Instruction Following | 1079 | 1357 |
| IFEval | — | 95.1% |
Long Context Grok-3 mini leads
DeepSeek LLM 67B: 33.1 (#265), Grok-3 mini: 41.0 (#147)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| LMArena Longer Query | 1092 | 1372 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference Grok-3 mini leads
DeepSeek LLM 67B: 31.6 (#282), Grok-3 mini: 52.5 (#169)
| Benchmark | DeepSeek LLM 67B | Grok-3 mini |
|---|---|---|
| LMArena Text | 1105 | 1370 |
| LMArena Creative Writing | 1067 | 1342 |
| LMArena Multi-Turn | 1082 | 1355 |
| Short-Story Creative Writing | — | 73.5% |
| WildBench | — | 65.1% |
Frequently asked questions
Is DeepSeek LLM 67B better than Grok-3 mini?
Grok-3 mini is the stronger model overall, scoring 41.2 to 24.9 on the Noometry Index.
Is DeepSeek LLM 67B or Grok-3 mini better for coding?
Grok-3 mini scores higher on coding benchmarks: 40.8 versus 31.9 in the Noometry coding category.
How many benchmarks do DeepSeek LLM 67B and Grok-3 mini share?
14 benchmarks have published results for both models. DeepSeek LLM 67B has 15 scored results on Noometry and Grok-3 mini has 35.