Model comparison
Deepseek Coder v2 vs Granite 4.2 30b
Granite 4.2 30b is the stronger model overall, scoring 41.8 to 35.9 on the Noometry Index.
Last verified . 11 shared benchmarks.
DeepSeekSummary
- They share 11 benchmarks with published results for both. Deepseek Coder v2 scores higher in 0 categories and Granite 4.2 30b in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Granite 4.2 30b leads 53.8 to 38.2.
Side by side
| Deepseek Coder v2 | Granite 4.2 30b | |
|---|---|---|
| Provider | DeepSeek | IBM |
| Noometry Index | 35.9 | 41.8 |
| Released | 2024-06-17 | — |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 24 | 11 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Granite 4.2 30b leads
Deepseek Coder v2: 38.1 (#183), Granite 4.2 30b: 41.0 (#126)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Coding | 1251 | 1396 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Reasoning Granite 4.2 30b leads
Deepseek Coder v2: 23.6 (#176), Granite 4.2 30b: 27.8 (#112)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1374 |
| WinoGrande | 83.7% | — |
Math Not comparable
Deepseek Coder v2: 34.9 (#190), Granite 4.2 30b: —
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Math | 1241 | — |
| GSM8K | 94.5% | — |
Knowledge Granite 4.2 30b leads
Deepseek Coder v2: 32.3 (#212), Granite 4.2 30b: 39.1 (#138)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Expert | 1181 | 1406 |
| ARC (AI2) Challenge | 64.3% | — |
Multilingual Granite 4.2 30b leads
Deepseek Coder v2: 36.3 (#240), Granite 4.2 30b: 47.3 (#151)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Non-English | 1182 | 1340 |
| LMArena Chinese | 1201 | 1414 |
| LMArena Russian | 1188 | 1343 |
| LMArena French | 1185 | — |
| LMArena German | 1164 | — |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
| LMArena Spanish | 1153 | — |
Instruction Following Granite 4.2 30b leads
Deepseek Coder v2: 61.7 (#242), Granite 4.2 30b: 71.2 (#155)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Instruction Following | 1180 | 1347 |
Long Context Granite 4.2 30b leads
Deepseek Coder v2: 37.0 (#224), Granite 4.2 30b: 41.4 (#140)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Longer Query | 1219 | 1359 |
Writing & Preference Granite 4.2 30b leads
Deepseek Coder v2: 38.2 (#253), Granite 4.2 30b: 53.8 (#156)
| Benchmark | Deepseek Coder v2 | Granite 4.2 30b |
|---|---|---|
| LMArena Text | 1191 | 1361 |
| LMArena Creative Writing | 1120 | 1288 |
| LMArena Multi-Turn | 1177 | 1339 |
Frequently asked questions
Is Deepseek Coder v2 better than Granite 4.2 30b?
Granite 4.2 30b is the stronger model overall, scoring 41.8 to 35.9 on the Noometry Index.
Is Deepseek Coder v2 or Granite 4.2 30b better for coding?
Granite 4.2 30b scores higher on coding benchmarks: 41.0 versus 38.1 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and Granite 4.2 30b share?
11 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and Granite 4.2 30b has 11.