Model comparison
Llama 3.1-405B vs Llama 3-8B
Llama 3.1-405B is the stronger model overall, scoring 30.4 to 25.3 on the Noometry Index.
Last verified . 27 shared benchmarks.
MetaSummary
- They share 27 benchmarks with published results for both. Llama 3.1-405B scores higher in 8 categories and Llama 3-8B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 3.1-405B leads 30.4 to 7.8.
- The biggest single-benchmark swing is MATH Level 5: 49.8% for Llama 3.1-405B and 6.1% for Llama 3-8B.
Side by side
| Llama 3.1-405B | Llama 3-8B | |
|---|---|---|
| Provider | Meta | Meta |
| Noometry Index | 30.4 | 25.3 |
| Released | 2024-07-23 | 2024-04-18 |
| Weights | Open | Open |
| Context window | — | 8K |
| Max output | — | 8K |
| Input $ / M tokens | — | $0.14 |
| Output $ / M tokens | — | $0.14 |
| Results tracked | 42 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3.1-405B leads
Llama 3.1-405B: 32.7 (#257), Llama 3-8B: 30.5 (#285)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| LMArena Coding | 1291 | 1152 |
| WeirdML | 21.4% | — |
| BigCodeBench Instruct | — | 31.9% |
| BigCodeBench Complete | — | 36.9% |
| HumanEval+ | — | 56.7% |
| MBPP+ | — | 54.8% |
Agentic & Tool Use Not comparable
Llama 3.1-405B: 20.9 (#131), Llama 3-8B: —
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| TheAgentCompany | 7.4% | — |
| Cybench | 7.5% | — |
Reasoning Llama 3.1-405B leads
Llama 3.1-405B: 16.4 (#293), Llama 3-8B: 13.8 (#321)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| LMArena Hard Prompts | 1270 | 1133 |
| DTBench | 61.4% | 43.9% |
| Epoch Capabilities Index | 128.76 | 116.33 |
| ForecastBench | 59.9 | 58.6 |
| WinoGrande | 89.2% | 75.7% |
| SimpleBench | 23% | — |
| Kagi LLM Benchmark | 45% | — |
| Chess Puzzles | — | 0% |
| Adversarial NLI | — | 57.3% |
| BIG-Bench Hard | 82.9% | — |
| HellaSwag | 89.2% | — |
| PIQA | 85.9% | — |
Math Llama 3.1-405B leads
Llama 3.1-405B: 17.3 (#283), Llama 3-8B: 8.3 (#317)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 9.7% | 1.9% |
| LMArena Math | 1281 | 1151 |
| MATH Level 5 | 49.8% | 6.1% |
| Omni-MATH | 24.9% | — |
Knowledge Llama 3.1-405B leads
Llama 3.1-405B: 30.4 (#221), Llama 3-8B: 7.8 (#303)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| GPQA Diamond | 50.9% | 26.1% |
| LMArena Expert | 1243 | 1113 |
| ARC (AI2) Challenge | 95.3% | 82.8% |
| MMLU | 84.5% | 68.8% |
| TriviaQA | 82.7% | 67.7% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 17.6% | — |
| GPQA (HELM) | 52.2% | — |
| OpenBookQA | — | 82.6% |
Multilingual Llama 3.1-405B leads
Llama 3.1-405B: 40.9 (#209), Llama 3-8B: 31.0 (#256)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 1248 | 1098 |
| LMArena Chinese | 1242 | 1076 |
| LMArena French | 1278 | 1158 |
| LMArena German | 1252 | 1104 |
| LMArena Japanese | 1208 | 967 |
| LMArena Korean | 1184 | 1003 |
| LMArena Russian | 1265 | 1109 |
| LMArena Spanish | 1259 | 1172 |
Instruction Following Llama 3.1-405B leads
Llama 3.1-405B: 66.1 (#209), Llama 3-8B: 58.7 (#255)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1259 | 1127 |
| IFEval | 81.1% | — |
Long Context Llama 3.1-405B leads
Llama 3.1-405B: 38.4 (#192), Llama 3-8B: 34.2 (#246)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1266 | 1128 |
Writing & Preference Llama 3.1-405B leads
Llama 3.1-405B: 39.1 (#246), Llama 3-8B: 37.7 (#251)
| Benchmark | Llama 3.1-405B | Llama 3-8B |
|---|---|---|
| LMArena Text | 1284 | 1166 |
| LMArena Creative Writing | 1262 | 1150 |
| LMArena Multi-Turn | 1297 | 1152 |
| EQ-Bench Creative Writing | 870 | — |
| WildBench | 78.3% | — |
Frequently asked questions
Is Llama 3.1-405B better than Llama 3-8B?
Llama 3.1-405B is the stronger model overall, scoring 30.4 to 25.3 on the Noometry Index.
Is Llama 3.1-405B or Llama 3-8B better for coding?
Llama 3.1-405B scores higher on coding benchmarks: 32.7 versus 30.5 in the Noometry coding category.
How many benchmarks do Llama 3.1-405B and Llama 3-8B share?
27 benchmarks have published results for both models. Llama 3.1-405B has 42 scored results on Noometry and Llama 3-8B has 34.