Model comparison
DeepSeek-V3.2-Speciale vs Grok-3 mini
Grok-3 mini is the stronger model overall, scoring 41.0 to 39.1 on the Noometry Index.
Last verified . 1 shared benchmarks.
DeepSeekSummary
- They share 1 benchmark with published results for both. DeepSeek-V3.2-Speciale scores higher in 1 category and Grok-3 mini in 2 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V3.2-Speciale leads 32.1 to 13.8.
- DeepSeek-V3.2-Speciale has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.2-Speciale | Grok-3 mini | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 39.1 | 41.0 |
| Released | 2025-12-01 | 2025-04-09 |
| Weights | Open | Proprietary |
| Context window | 128K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.58 | — |
| Output $ / M tokens | $1.68 | — |
| Results tracked | 3 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-V3.2-Speciale: 39.9 (#136), Grok-3 mini: 40.4 (#126)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| WeirdML | 46.7% | 42.6% |
| Aider Polyglot | — | 49.3% |
| LMArena Coding | — | 1380 |
Reasoning DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 32.1 (#65), Grok-3 mini: 13.8 (#323)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| ARC-AGI-2 | — | 0.4% |
| SimpleBench | 52.6% | — |
| Kagi LLM Benchmark | — | 61.3% |
| ARC-AGI-1 | — | 16.5% |
| LMArena Hard Prompts | — | 1374 |
| Epoch Capabilities Index | — | 140.35 |
Math Not comparable
DeepSeek-V3.2-Speciale: —, Grok-3 mini: 40.3 (#78)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 77.8% |
| Omni-MATH | — | 31.8% |
| LMArena Math | — | 1386 |
| MATH Level 5 | — | 90.9% |
| FrontierMath (Feb 2025 set) | — | 5.9% |
Knowledge Not comparable
DeepSeek-V3.2-Speciale: —, Grok-3 mini: 46.3 (#75)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| GPQA Diamond | — | 76.3% |
| MMLU-Pro | — | 79.9% |
| Confabulations | — | 10.8% |
| GPQA (HELM) | — | 67.5% |
| LMArena Expert | — | 1395 |
Multilingual Not comparable
DeepSeek-V3.2-Speciale: —, Grok-3 mini: 48.3 (#141)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| LMArena Non-English | — | 1352 |
| LMArena Chinese | — | 1387 |
| LMArena French | — | 1357 |
| LMArena German | — | 1350 |
| LMArena Japanese | — | 1343 |
| LMArena Korean | — | 1336 |
| LMArena Russian | — | 1352 |
| LMArena Spanish | — | 1379 |
Instruction Following Not comparable
DeepSeek-V3.2-Speciale: —, Grok-3 mini: 78.7 (#8)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| IFEval | — | 95.1% |
| LMArena Instruction Following | — | 1356 |
Long Context Not comparable
DeepSeek-V3.2-Speciale: —, Grok-3 mini: 41.0 (#142)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| Fiction.LiveBench | — | 66.7% |
| LMArena Longer Query | — | 1372 |
Writing & Preference Grok-3 mini leads
DeepSeek-V3.2-Speciale: 46.1 (#217), Grok-3 mini: 52.7 (#163)
| Benchmark | DeepSeek-V3.2-Speciale | Grok-3 mini |
|---|---|---|
| LMArena Text | — | 1370 |
| LMArena Creative Writing | — | 1342 |
| Short-Story Creative Writing | — | 73.5% |
| EQ-Bench Creative Writing | 1276 | — |
| WildBench | — | 65.1% |
| LMArena Multi-Turn | — | 1355 |
Frequently asked questions
Is DeepSeek-V3.2-Speciale better than Grok-3 mini?
Grok-3 mini is the stronger model overall, scoring 41.0 to 39.1 on the Noometry Index.
Is DeepSeek-V3.2-Speciale or Grok-3 mini better for coding?
They score almost the same on coding (39.9 vs 40.4); test both on your own repository before choosing.
How many benchmarks do DeepSeek-V3.2-Speciale and Grok-3 mini share?
1 benchmark has published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and Grok-3 mini has 35.