Model comparison
Codestral vs Granite 4.0 H Small
Granite 4.0 H Small is the stronger model overall, scoring 36.5 to 30.6 on the Noometry Index.
Last verified . 0 shared benchmarks.
Mistral AISummary
- The widest gap is in coding, where Granite 4.0 H Small leads 36.4 to 27.3.
- Granite 4.0 H Small has downloadable open weights; the other is API-only.
Side by side
| Codestral | Granite 4.0 H Small | |
|---|---|---|
| Provider | Mistral AI | IBM |
| Noometry Index | 30.6 | 36.5 |
| Released | 2024-05-29 | — |
| Weights | Proprietary | Open |
| Context window | 256K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.30 | — |
| Output $ / M tokens | $0.90 | — |
| Results tracked | 7 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Granite 4.0 H Small leads
Codestral: 27.3 (#321), Granite 4.0 H Small: 36.4 (#209)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| Aider Polyglot | 11.1% | — |
| BigCodeBench Instruct | 41.8% | — |
| LMArena Coding | — | 1249 |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | 137.78 | — |
| HumanEval+ | 73.8% | — |
| MBPP+ | 61.9% | — |
Reasoning Granite 4.0 H Small leads
Codestral: 19.8 (#251), Granite 4.0 H Small: 24.4 (#163)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| Kagi LLM Benchmark | 32.5% | — |
| LMArena Hard Prompts | — | 1240 |
Math Not comparable
Codestral: —, Granite 4.0 H Small: 31.4 (#223)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| Omni-MATH | — | 29.6% |
| LMArena Math | — | 1247 |
Knowledge Not comparable
Codestral: —, Granite 4.0 H Small: 32.0 (#215)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| MMLU-Pro | — | 56.9% |
| Vectara Hallucination Rate | — | 5.2% |
| GPQA (HELM) | — | 38.3% |
| LMArena Expert | — | 1251 |
Multilingual Not comparable
Codestral: —, Granite 4.0 H Small: 38.6 (#228)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| LMArena Non-English | — | 1216 |
| LMArena Chinese | — | 1249 |
| LMArena Russian | — | 1201 |
| LMArena Spanish | — | 1258 |
Instruction Following Not comparable
Codestral: —, Granite 4.0 H Small: 68.7 (#183)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| IFEval | — | 89% |
| LMArena Instruction Following | — | 1222 |
Long Context Not comparable
Codestral: —, Granite 4.0 H Small: 37.7 (#213)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| LMArena Longer Query | — | 1242 |
Writing & Preference Not comparable
Codestral: —, Granite 4.0 H Small: 43.8 (#227)
| Benchmark | Codestral | Granite 4.0 H Small |
|---|---|---|
| LMArena Text | — | 1241 |
| LMArena Creative Writing | — | 1211 |
| WildBench | — | 73.9% |
| LMArena Multi-Turn | — | 1242 |
Frequently asked questions
Is Codestral better than Granite 4.0 H Small?
Granite 4.0 H Small is the stronger model overall, scoring 36.5 to 30.6 on the Noometry Index.
Is Codestral or Granite 4.0 H Small better for coding?
Granite 4.0 H Small scores higher on coding benchmarks: 36.4 versus 27.3 in the Noometry coding category.
How many benchmarks do Codestral and Granite 4.0 H Small share?
0 benchmarks have published results for both models. Codestral has 7 scored results on Noometry and Granite 4.0 H Small has 19.