Model comparison
Granite 4.0 Micro vs Llama 3.1-405B
Llama 3.1-405B is the stronger model overall, scoring 30.4 to 28.8 on the Noometry Index.
Last verified . 7 shared benchmarks.
MetaSummary
- They share 7 benchmarks with published results for both. Granite 4.0 Micro scores higher in 3 categories and Llama 3.1-405B in 2 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 3.1-405B leads 30.4 to 9.9.
- The biggest single-benchmark swing is MMLU-Pro: 39.5% for Granite 4.0 Micro and 72.3% for Llama 3.1-405B.
Side by side
| Granite 4.0 Micro | Llama 3.1-405B | |
|---|---|---|
| Provider | IBM | Meta |
| Noometry Index | 28.8 | 30.4 |
| Released | 2025-10-02 | 2024-07-23 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 118K | — |
| Input $ / M tokens | $0.017 | — |
| Output $ / M tokens | $0.11 | — |
| Results tracked | 8 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Granite 4.0 Micro: —, Llama 3.1-405B: 32.7 (#257)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| WeirdML | — | 21.4% |
| LMArena Coding | — | 1291 |
Agentic & Tool Use Not comparable
Granite 4.0 Micro: —, Llama 3.1-405B: 20.9 (#131)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| TheAgentCompany | — | 7.4% |
| Cybench | — | 7.5% |
Reasoning Granite 4.0 Micro leads
Granite 4.0 Micro: 18.5 (#261), Llama 3.1-405B: 16.4 (#293)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| SimpleBench | — | 23% |
| Kagi LLM Benchmark | — | 45% |
| Chess Puzzles | 0% | — |
| LMArena Hard Prompts | — | 1270 |
| DTBench | — | 61.4% |
| BIG-Bench Hard | — | 82.9% |
| Epoch Capabilities Index | — | 128.76 |
| ForecastBench | — | 59.9 |
| HellaSwag | — | 89.2% |
| PIQA | — | 85.9% |
| WinoGrande | — | 89.2% |
Math Llama 3.1-405B leads
Granite 4.0 Micro: 11.2 (#301), Llama 3.1-405B: 17.3 (#283)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.8% | 9.7% |
| Omni-MATH | 20.9% | 24.9% |
| LMArena Math | — | 1281 |
| MATH Level 5 | — | 49.8% |
Knowledge Llama 3.1-405B leads
Granite 4.0 Micro: 9.9 (#299), Llama 3.1-405B: 30.4 (#221)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| GPQA Diamond | 28.3% | 50.9% |
| MMLU-Pro | 39.5% | 72.3% |
| GPQA (HELM) | 30.7% | 52.2% |
| Confabulations | — | 17.6% |
| LMArena Expert | — | 1243 |
| ARC (AI2) Challenge | — | 95.3% |
| MMLU | — | 84.5% |
| TriviaQA | — | 82.7% |
Multilingual Not comparable
Granite 4.0 Micro: —, Llama 3.1-405B: 40.9 (#209)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| LMArena Non-English | — | 1248 |
| LMArena Chinese | — | 1242 |
| LMArena French | — | 1278 |
| LMArena German | — | 1252 |
| LMArena Japanese | — | 1208 |
| LMArena Korean | — | 1184 |
| LMArena Russian | — | 1265 |
| LMArena Spanish | — | 1259 |
Instruction Following Granite 4.0 Micro leads
Granite 4.0 Micro: 70.0 (#164), Llama 3.1-405B: 66.1 (#209)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| IFEval | 84.9% | 81.1% |
| LMArena Instruction Following | — | 1259 |
Long Context Not comparable
Granite 4.0 Micro: —, Llama 3.1-405B: 38.4 (#192)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| LMArena Longer Query | — | 1266 |
Writing & Preference Granite 4.0 Micro leads
Granite 4.0 Micro: 46.7 (#212), Llama 3.1-405B: 39.1 (#246)
| Benchmark | Granite 4.0 Micro | Llama 3.1-405B |
|---|---|---|
| WildBench | 67% | 78.3% |
| LMArena Text | — | 1284 |
| LMArena Creative Writing | — | 1262 |
| EQ-Bench Creative Writing | — | 870 |
| LMArena Multi-Turn | — | 1297 |
Frequently asked questions
Is Granite 4.0 Micro better than Llama 3.1-405B?
Llama 3.1-405B is the stronger model overall, scoring 30.4 to 28.8 on the Noometry Index.
How many benchmarks do Granite 4.0 Micro and Llama 3.1-405B share?
7 benchmarks have published results for both models. Granite 4.0 Micro has 8 scored results on Noometry and Llama 3.1-405B has 42.