Model comparison
Gemma 1.1 2b IT vs Llama 2-70B
Gemma 1.1 2b IT is the stronger model overall, scoring 29.3 to 24.4 on the Noometry Index.
Last verified . 14 shared benchmarks.
Google
MetaSummary
- They share 14 benchmarks with published results for both. Gemma 1.1 2b IT scores higher in 3 categories and Llama 2-70B in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemma 1.1 2b IT leads 30.8 to 8.1.
Side by side
| Gemma 1.1 2b IT | Llama 2-70B | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 29.3 | 24.4 |
| Released | — | 2023-07-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 16 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-70B leads
Gemma 1.1 2b IT: 30.1 (#299), Llama 2-70B: 31.4 (#286)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1034 | 1079 |
| HumanEval+ | 17.7% | — |
| MBPP+ | 23.3% | — |
Reasoning Gemma 1.1 2b IT leads
Gemma 1.1 2b IT: 19.1 (#270), Llama 2-70B: 14.4 (#325)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1005 | 1073 |
| DTBench | — | 41.6% |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| Epoch Capabilities Index | — | 113.79 |
| ForecastBench | — | 51.4 |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| PIQA | — | 82.8% |
| WinoGrande | — | 80.2% |
Math Gemma 1.1 2b IT leads
Gemma 1.1 2b IT: 30.8 (#232), Llama 2-70B: 8.1 (#326)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Math | 1047 | 1091 |
| OTIS Mock AIME 2024-2025 | — | 0% |
| MATH Level 5 | — | 3.3% |
| GSM8K | — | 69.6% |
Knowledge Gemma 1.1 2b IT leads
Gemma 1.1 2b IT: 26.5 (#258), Llama 2-70B: 7.4 (#310)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Expert | 970 | 1039 |
| GPQA Diamond | — | 26.3% |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| MMLU | — | 69.9% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |
Multilingual Llama 2-70B leads
Gemma 1.1 2b IT: 24.6 (#289), Llama 2-70B: 27.7 (#274)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 988 | 1045 |
| LMArena Chinese | 1012 | 995 |
| LMArena German | 944 | 1041 |
| LMArena Korean | 899 | 964 |
| LMArena Russian | 990 | 1083 |
| LMArena French | — | 1090 |
| LMArena Japanese | — | 927 |
| LMArena Spanish | — | 1143 |
Instruction Following Llama 2-70B leads
Gemma 1.1 2b IT: 49.9 (#299), Llama 2-70B: 54.9 (#278)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 992 | 1071 |
Long Context Llama 2-70B leads
Gemma 1.1 2b IT: 30.6 (#286), Llama 2-70B: 32.3 (#270)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1003 | 1062 |
Writing & Preference Llama 2-70B leads
Gemma 1.1 2b IT: 25.1 (#306), Llama 2-70B: 32.3 (#279)
| Benchmark | Gemma 1.1 2b IT | Llama 2-70B |
|---|---|---|
| LMArena Text | 1022 | 1115 |
| LMArena Creative Writing | 998 | 1075 |
| LMArena Multi-Turn | 959 | 1088 |
Frequently asked questions
Is Gemma 1.1 2b IT better than Llama 2-70B?
Gemma 1.1 2b IT is the stronger model overall, scoring 29.3 to 24.4 on the Noometry Index.
Is Gemma 1.1 2b IT or Llama 2-70B better for coding?
Llama 2-70B scores higher on coding benchmarks: 31.4 versus 30.1 in the Noometry coding category.
How many benchmarks do Gemma 1.1 2b IT and Llama 2-70B share?
14 benchmarks have published results for both models. Gemma 1.1 2b IT has 16 scored results on Noometry and Llama 2-70B has 35.