Model comparison
Llama 2-70B vs Llama 3.1-405B
Llama 3.1-405B is the stronger model overall, scoring 30.7 to 24.4 on the Noometry Index.
Last verified . 30 shared benchmarks.
MetaSummary
- They share 30 benchmarks with published results for both. Llama 2-70B scores higher in 0 categories and Llama 3.1-405B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 3.1-405B leads 30.4 to 7.4.
- The biggest single-benchmark swing is MATH Level 5: 3.3% for Llama 2-70B and 49.8% for Llama 3.1-405B.
Side by side
| Llama 2-70B | Llama 3.1-405B | |
|---|---|---|
| Provider | Meta | Meta |
| Noometry Index | 24.4 | 30.7 |
| Released | 2023-07-18 | 2024-07-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 35 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3.1-405B leads
Llama 2-70B: 31.4 (#286), Llama 3.1-405B: 33.1 (#262)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| LMArena Coding | 1079 | 1291 |
| WeirdML | — | 21.4% |
Agentic & Tool Use Not comparable
Llama 2-70B: —, Llama 3.1-405B: 21.0 (#140)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| TheAgentCompany | — | 7.4% |
| Cybench | — | 7.5% |
Reasoning Llama 3.1-405B leads
Llama 2-70B: 14.4 (#325), Llama 3.1-405B: 16.8 (#300)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1269 |
| DTBench | 41.6% | 61.4% |
| BIG-Bench Hard | 64.9% | 82.9% |
| Epoch Capabilities Index | 113.79 | 128.75 |
| ForecastBench | 51.4 | 59.9 |
| HellaSwag | 85.3% | 89.2% |
| PIQA | 82.8% | 85.9% |
| WinoGrande | 80.2% | 89.2% |
| SimpleBench | — | 23% |
| Kagi LLM Benchmark | — | 45% |
| CommonsenseQA 2.0 | 50% | — |
| LAMBADA | 78.9% | — |
Math Llama 3.1-405B leads
Llama 2-70B: 8.1 (#326), Llama 3.1-405B: 18.4 (#290)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0% | 9.7% |
| LMArena Math | 1091 | 1281 |
| MATH Level 5 | 3.3% | 49.8% |
| Omni-MATH | — | 24.9% |
| GSM8K | 69.6% | — |
Knowledge Llama 3.1-405B leads
Llama 2-70B: 7.4 (#310), Llama 3.1-405B: 30.4 (#227)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| GPQA Diamond | 26.3% | 50.9% |
| LMArena Expert | 1039 | 1243 |
| ARC (AI2) Challenge | 78.3% | 95.3% |
| MMLU | 69.9% | 84.5% |
| TriviaQA | 87.6% | 82.7% |
| MMLU-Pro | — | 72.3% |
| Confabulations | — | 17.6% |
| GPQA (HELM) | — | 52.2% |
| BoolQ | 88.6% | — |
| OpenBookQA | 60.2% | — |
Multilingual Llama 3.1-405B leads
Llama 2-70B: 27.7 (#274), Llama 3.1-405B: 40.7 (#214)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| LMArena Non-English | 1045 | 1248 |
| LMArena Chinese | 995 | 1242 |
| LMArena French | 1090 | 1279 |
| LMArena German | 1041 | 1252 |
| LMArena Japanese | 927 | 1208 |
| LMArena Korean | 964 | 1184 |
| LMArena Russian | 1083 | 1265 |
| LMArena Spanish | 1143 | 1260 |
Instruction Following Llama 3.1-405B leads
Llama 2-70B: 54.9 (#278), Llama 3.1-405B: 65.9 (#214)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1259 |
| IFEval | — | 81.1% |
Long Context Llama 3.1-405B leads
Llama 2-70B: 32.3 (#270), Llama 3.1-405B: 38.4 (#197)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| LMArena Longer Query | 1062 | 1266 |
Writing & Preference Llama 3.1-405B leads
Llama 2-70B: 32.3 (#279), Llama 3.1-405B: 38.9 (#251)
| Benchmark | Llama 2-70B | Llama 3.1-405B |
|---|---|---|
| LMArena Text | 1115 | 1284 |
| LMArena Creative Writing | 1075 | 1262 |
| LMArena Multi-Turn | 1088 | 1297 |
| EQ-Bench Creative Writing | — | 870 |
| WildBench | — | 78.3% |
Frequently asked questions
Is Llama 2-70B better than Llama 3.1-405B?
Llama 3.1-405B is the stronger model overall, scoring 30.7 to 24.4 on the Noometry Index.
Is Llama 2-70B or Llama 3.1-405B better for coding?
Llama 3.1-405B scores higher on coding benchmarks: 33.1 versus 31.4 in the Noometry coding category.
How many benchmarks do Llama 2-70B and Llama 3.1-405B share?
30 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and Llama 3.1-405B has 42.