Model comparison
Gemma 2B vs Llama 13b
Gemma 2B is the stronger model overall, scoring 29.6 to 24.4 on the Noometry Index.
Last verified . 18 shared benchmarks.
Google
MetaSummary
- They share 18 benchmarks with published results for both. Gemma 2B scores higher in 6 categories and Llama 13b in 0 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Gemma 2B leads 48.5 to 36.7.
Side by side
| Gemma 2B | Llama 13b | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 29.6 | 24.4 |
| Released | 2024-02-21 | 2023-02-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 23 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 2B leads
Gemma 2B: 29.4 (#305), Llama 13b: 21.4 (#337)
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Coding | 1010 | 683 |
| HumanEval+ | 20.7% | — |
| MBPP+ | 34.1% | — |
Reasoning Gemma 2B leads
Gemma 2B: 18.8 (#275), Llama 13b: 14.0 (#329)
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Hard Prompts | 989 | 728 |
| BIG-Bench Hard | 35.2% | 37.9% |
| Epoch Capabilities Index | 94.2 | 100.58 |
| HellaSwag | 71.4% | 79.2% |
| PIQA | 77.3% | 80.1% |
| WinoGrande | 65.4% | 73% |
| LAMBADA | — | 75.2% |
Math Gemma 2B leads
Gemma 2B: 30.0 (#239), Llama 13b: 26.7 (#256)
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Math | 1009 | 838 |
| GSM8K | 17.7% | 20.6% |
Knowledge Not comparable
Gemma 2B: —, Llama 13b: —
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| ARC (AI2) Challenge | 42.1% | 52.7% |
| BoolQ | 69.4% | 78.7% |
| MMLU | 42.3% | 47.7% |
| TriviaQA | 53.2% | 77.9% |
| OpenBookQA | — | 56.4% |
Multimodal Not comparable
Gemma 2B: —, Llama 13b: —
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| ScienceQA | — | 43.3% |
Multilingual Gemma 2B leads
Gemma 2B: 23.0 (#294), Llama 13b: 16.6 (#297)
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Non-English | 958 | 819 |
| LMArena Chinese | 986 | — |
| LMArena Russian | 937 | — |
Instruction Following Gemma 2B leads
Gemma 2B: 48.5 (#302), Llama 13b: 36.7 (#305)
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Instruction Following | 970 | 781 |
Long Context Not comparable
Gemma 2B: 29.9 (#291), Llama 13b: —
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Longer Query | 981 | — |
Writing & Preference Gemma 2B leads
Gemma 2B: 24.0 (#308), Llama 13b: 13.8 (#312)
| Benchmark | Gemma 2B | Llama 13b |
|---|---|---|
| LMArena Text | 1002 | 834 |
| LMArena Creative Writing | 987 | 794 |
| LMArena Multi-Turn | 945 | 753 |
Frequently asked questions
Is Gemma 2B better than Llama 13b?
Gemma 2B is the stronger model overall, scoring 29.6 to 24.4 on the Noometry Index.
Is Gemma 2B or Llama 13b better for coding?
Gemma 2B scores higher on coding benchmarks: 29.4 versus 21.4 in the Noometry coding category.
How many benchmarks do Gemma 2B and Llama 13b share?
18 benchmarks have published results for both models. Gemma 2B has 23 scored results on Noometry and Llama 13b has 21.