Model comparison
Llama 2-70B vs Llama 2-7B
Llama 2-7B is the stronger model overall, scoring 28.6 to 24.1 on the Noometry Index.
Last verified . 27 shared benchmarks.
MetaSummary
- They share 27 benchmarks with published results for both. Llama 2-70B scores higher in 5 categories and Llama 2-7B in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Llama 2-7B leads 28.2 to 7.5.
Side by side
| Llama 2-70B | Llama 2-7B | |
|---|---|---|
| Provider | Meta | Meta |
| Noometry Index | 24.1 | 28.6 |
| Released | 2023-07-18 | 2023-07-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 35 | 29 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-70B leads
Llama 2-70B: 30.8 (#281), Llama 2-7B: 28.7 (#302)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Coding | 1079 | 1002 |
Reasoning Llama 2-7B leads
Llama 2-70B: 13.7 (#324), Llama 2-7B: 15.1 (#309)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1009 |
| BIG-Bench Hard | 64.9% | 39.2% |
| Epoch Capabilities Index | 113.59 | 98.54 |
| HellaSwag | 85.3% | 77.2% |
| LAMBADA | 78.9% | 73.3% |
| PIQA | 82.8% | 78.8% |
| WinoGrande | 80.2% | 69.2% |
| Chess Puzzles | — | 0% |
| DTBench | 41.6% | — |
| CommonsenseQA 2.0 | 50% | — |
| ForecastBench | 51.4 | — |
Math Llama 2-7B leads
Llama 2-70B: 7.5 (#320), Llama 2-7B: 28.2 (#231)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Math | 1091 | 1042 |
| GSM8K | 69.6% | 16.7% |
| OTIS Mock AIME 2024-2025 | 0% | — |
| MATH Level 5 | 3.3% | — |
Knowledge Llama 2-7B leads
Llama 2-70B: 7.4 (#305), Llama 2-7B: 27.7 (#243)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Expert | 1039 | 1036 |
| ARC (AI2) Challenge | 78.3% | 45.9% |
| BoolQ | 88.6% | 77.9% |
| MMLU | 69.9% | 45.8% |
| OpenBookQA | 60.2% | 58.6% |
| TriviaQA | 87.6% | 73.7% |
| GPQA Diamond | 26.3% | — |
Multimodal Not comparable
Llama 2-70B: —, Llama 2-7B: —
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| ScienceQA | — | 43.1% |
Multilingual Llama 2-70B leads
Llama 2-70B: 27.9 (#269), Llama 2-7B: 23.9 (#288)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Non-English | 1045 | 973 |
| LMArena Chinese | 994 | 973 |
| LMArena French | 1089 | 969 |
| LMArena German | 1041 | 978 |
| LMArena Russian | 1083 | 995 |
| LMArena Spanish | 1142 | 1007 |
| LMArena Japanese | 927 | — |
| LMArena Korean | 964 | — |
Instruction Following Llama 2-70B leads
Llama 2-70B: 55.1 (#273), Llama 2-7B: 51.0 (#293)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1006 |
Long Context Llama 2-70B leads
Llama 2-70B: 32.2 (#265), Llama 2-7B: 30.4 (#282)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Longer Query | 1062 | 999 |
Writing & Preference Llama 2-70B leads
Llama 2-70B: 32.5 (#274), Llama 2-7B: 28.1 (#293)
| Benchmark | Llama 2-70B | Llama 2-7B |
|---|---|---|
| LMArena Text | 1115 | 1053 |
| LMArena Creative Writing | 1075 | 1033 |
| LMArena Multi-Turn | 1088 | 1029 |
Frequently asked questions
Is Llama 2-70B better than Llama 2-7B?
Llama 2-7B is the stronger model overall, scoring 28.6 to 24.1 on the Noometry Index.
Is Llama 2-70B or Llama 2-7B better for coding?
Llama 2-70B scores higher on coding benchmarks: 30.8 versus 28.7 in the Noometry coding category.
How many benchmarks do Llama 2-70B and Llama 2-7B share?
27 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and Llama 2-7B has 29.