Model comparison
Granite 4.0 H Small vs Qwen2.5-Coder-32B
Granite 4.0 H Small is the stronger model overall, scoring 36.5 to 33.4 on the Noometry Index.
Last verified . 12 shared benchmarks.
Alibaba (Qwen)Summary
- They share 12 benchmarks with published results for both. Granite 4.0 H Small scores higher in 5 categories and Qwen2.5-Coder-32B in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in coding, where Granite 4.0 H Small leads 36.4 to 22.6.
Side by side
| Granite 4.0 H Small | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | IBM | Alibaba (Qwen) |
| Noometry Index | 36.5 | 33.4 |
| Released | — | 2024-09-18 |
| Weights | Open | Open |
| Context window | — | 33K |
| Max output | — | 29K |
| Input $ / M tokens | — | $0.66 |
| Output $ / M tokens | — | $1 |
| Results tracked | 19 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Granite 4.0 H Small leads
Granite 4.0 H Small: 36.4 (#209), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Coding | 1249 | 1276 |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Reasoning Granite 4.0 H Small leads
Granite 4.0 H Small: 24.4 (#163), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1240 | 1251 |
| LiveBench Reasoning | — | 42.1% |
| LiveBench Data Analysis | — | 49.9% |
| Epoch Capabilities Index | — | 119.49 |
| HellaSwag | — | 83% |
| LiveBench | — | 46.2% |
| WinoGrande | — | 80.8% |
Math Qwen2.5-Coder-32B leads
Granite 4.0 H Small: 31.4 (#223), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1247 | 1251 |
| Omni-MATH | 29.6% | — |
| LiveBench Math | — | 46.6% |
| GSM8K | — | 93% |
Knowledge Qwen2.5-Coder-32B leads
Granite 4.0 H Small: 32.0 (#215), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1251 | 1221 |
| MMLU-Pro | 56.9% | — |
| Vectara Hallucination Rate | 5.2% | — |
| GPQA (HELM) | 38.3% | — |
| ARC (AI2) Challenge | — | 70.5% |
| MMLU | — | 79.1% |
Multilingual Too close to call
Granite 4.0 H Small: 38.6 (#228), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1216 | 1205 |
| LMArena Chinese | 1249 | 1222 |
| LMArena Russian | 1201 | 1228 |
| LMArena Spanish | 1258 | — |
Instruction Following Granite 4.0 H Small leads
Granite 4.0 H Small: 68.7 (#183), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1222 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
| IFEval | 89% | — |
Long Context Too close to call
Granite 4.0 H Small: 37.7 (#213), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1242 | 1251 |
Writing & Preference Granite 4.0 H Small leads
Granite 4.0 H Small: 43.8 (#227), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | Granite 4.0 H Small | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1241 | 1230 |
| LMArena Creative Writing | 1211 | 1174 |
| LMArena Multi-Turn | 1242 | 1222 |
| WildBench | 73.9% | — |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is Granite 4.0 H Small better than Qwen2.5-Coder-32B?
Granite 4.0 H Small is the stronger model overall, scoring 36.5 to 33.4 on the Noometry Index.
Is Granite 4.0 H Small or Qwen2.5-Coder-32B better for coding?
Granite 4.0 H Small scores higher on coding benchmarks: 36.4 versus 22.6 in the Noometry coding category.
How many benchmarks do Granite 4.0 H Small and Qwen2.5-Coder-32B share?
12 benchmarks have published results for both models. Granite 4.0 H Small has 19 scored results on Noometry and Qwen2.5-Coder-32B has 31.