Model comparison
Granite 4.0 Micro vs Llama 3-8B
Granite 4.0 Micro is the stronger model overall, scoring 29.0 to 25.5 on the Noometry Index.
Last verified . 3 shared benchmarks.
MetaSummary
- They share 3 benchmarks with published results for both. Granite 4.0 Micro scores higher in 5 categories and Llama 3-8B in 0 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Granite 4.0 Micro leads 69.9 to 58.4.
Side by side
| Granite 4.0 Micro | Llama 3-8B | |
|---|---|---|
| Provider | IBM | Meta |
| Noometry Index | 29.0 | 25.5 |
| Released | 2025-10-02 | 2024-04-18 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 118K | — |
| Input $ / M tokens | $0.017 | — |
| Output $ / M tokens | $0.11 | — |
| Results tracked | 8 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Granite 4.0 Micro: —, Llama 3-8B: 31.0 (#289)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| BigCodeBench Instruct | — | 31.9% |
| LMArena Coding | — | 1152 |
| BigCodeBench Complete | — | 36.9% |
| HumanEval+ | — | 56.7% |
| MBPP+ | — | 54.8% |
Reasoning Granite 4.0 Micro leads
Granite 4.0 Micro: 19.2 (#265), Llama 3-8B: 14.3 (#326)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | — | 1133 |
| DTBench | — | 43.9% |
| Adversarial NLI | — | 57.3% |
| Epoch Capabilities Index | — | 116.45 |
| ForecastBench | — | 58.6 |
| WinoGrande | — | 75.7% |
Math Granite 4.0 Micro leads
Granite 4.0 Micro: 12.0 (#307), Llama 3-8B: 8.8 (#323)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.8% | 1.9% |
| Omni-MATH | 20.9% | — |
| LMArena Math | — | 1151 |
| MATH Level 5 | — | 6.1% |
Knowledge Granite 4.0 Micro leads
Granite 4.0 Micro: 9.9 (#304), Llama 3-8B: 7.8 (#308)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| GPQA Diamond | 28.3% | 26.1% |
| MMLU-Pro | 39.5% | — |
| GPQA (HELM) | 30.7% | — |
| LMArena Expert | — | 1113 |
| ARC (AI2) Challenge | — | 82.8% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 82.6% |
| TriviaQA | — | 67.7% |
Multilingual Not comparable
Granite 4.0 Micro: —, Llama 3-8B: 30.8 (#261)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| LMArena Non-English | — | 1098 |
| LMArena Chinese | — | 1076 |
| LMArena French | — | 1159 |
| LMArena German | — | 1104 |
| LMArena Japanese | — | 967 |
| LMArena Korean | — | 1004 |
| LMArena Russian | — | 1109 |
| LMArena Spanish | — | 1173 |
Instruction Following Granite 4.0 Micro leads
Granite 4.0 Micro: 69.9 (#169), Llama 3-8B: 58.4 (#260)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| IFEval | 84.9% | — |
| LMArena Instruction Following | — | 1127 |
Long Context Not comparable
Granite 4.0 Micro: —, Llama 3-8B: 34.2 (#251)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | — | 1128 |
Writing & Preference Granite 4.0 Micro leads
Granite 4.0 Micro: 46.7 (#216), Llama 3-8B: 37.5 (#256)
| Benchmark | Granite 4.0 Micro | Llama 3-8B |
|---|---|---|
| LMArena Text | — | 1166 |
| LMArena Creative Writing | — | 1150 |
| WildBench | 67% | — |
| LMArena Multi-Turn | — | 1152 |
Frequently asked questions
Is Granite 4.0 Micro better than Llama 3-8B?
Granite 4.0 Micro is the stronger model overall, scoring 29.0 to 25.5 on the Noometry Index.
How many benchmarks do Granite 4.0 Micro and Llama 3-8B share?
3 benchmarks have published results for both models. Granite 4.0 Micro has 8 scored results on Noometry and Llama 3-8B has 34.