Model comparison
DeepSeek-V3.2-Speciale vs Granite 4.0 H Small
DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 36.5 on the Noometry Index.
Last verified . 0 shared benchmarks.
DeepSeekSummary
- The widest gap is in reasoning, where DeepSeek-V3.2-Speciale leads 32.9 to 24.4.
Side by side
| DeepSeek-V3.2-Speciale | Granite 4.0 H Small | |
|---|---|---|
| Provider | DeepSeek | IBM |
| Noometry Index | 39.7 | 36.5 |
| Released | 2025-12-01 | — |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.58 | — |
| Output $ / M tokens | $1.68 | — |
| Results tracked | 3 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 40.4 (#140), Granite 4.0 H Small: 36.4 (#209)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| WeirdML | 46.7% | — |
| LMArena Coding | — | 1249 |
Reasoning DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 32.9 (#73), Granite 4.0 H Small: 24.4 (#163)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| SimpleBench | 52.6% | — |
| LMArena Hard Prompts | — | 1240 |
Math Not comparable
DeepSeek-V3.2-Speciale: —, Granite 4.0 H Small: 31.4 (#223)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| Omni-MATH | — | 29.6% |
| LMArena Math | — | 1247 |
Knowledge Not comparable
DeepSeek-V3.2-Speciale: —, Granite 4.0 H Small: 32.0 (#215)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| MMLU-Pro | — | 56.9% |
| Vectara Hallucination Rate | — | 5.2% |
| GPQA (HELM) | — | 38.3% |
| LMArena Expert | — | 1251 |
Multilingual Not comparable
DeepSeek-V3.2-Speciale: —, Granite 4.0 H Small: 38.6 (#228)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| LMArena Non-English | — | 1216 |
| LMArena Chinese | — | 1249 |
| LMArena Russian | — | 1201 |
| LMArena Spanish | — | 1258 |
Instruction Following Not comparable
DeepSeek-V3.2-Speciale: —, Granite 4.0 H Small: 68.7 (#183)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| IFEval | — | 89% |
| LMArena Instruction Following | — | 1222 |
Long Context Not comparable
DeepSeek-V3.2-Speciale: —, Granite 4.0 H Small: 37.7 (#213)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| LMArena Longer Query | — | 1242 |
Writing & Preference DeepSeek-V3.2-Speciale leads
DeepSeek-V3.2-Speciale: 46.0 (#222), Granite 4.0 H Small: 43.8 (#227)
| Benchmark | DeepSeek-V3.2-Speciale | Granite 4.0 H Small |
|---|---|---|
| LMArena Text | — | 1241 |
| LMArena Creative Writing | — | 1211 |
| EQ-Bench Creative Writing | 1276 | — |
| WildBench | — | 73.9% |
| LMArena Multi-Turn | — | 1242 |
Frequently asked questions
Is DeepSeek-V3.2-Speciale better than Granite 4.0 H Small?
DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 36.5 on the Noometry Index.
Is DeepSeek-V3.2-Speciale or Granite 4.0 H Small better for coding?
DeepSeek-V3.2-Speciale scores higher on coding benchmarks: 40.4 versus 36.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.2-Speciale and Granite 4.0 H Small share?
0 benchmarks have published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and Granite 4.0 H Small has 19.