Model comparison
Gemma 7B vs Llama 3-8B
Gemma 7B is the stronger model overall, scoring 29.5 to 25.3 on the Noometry Index.
Last verified . 22 shared benchmarks.
Google
MetaSummary
- They share 22 benchmarks with published results for both. Gemma 7B scores higher in 3 categories and Llama 3-8B in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemma 7B leads 28.7 to 8.3.
Side by side
| Gemma 7B | Llama 3-8B | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 29.5 | 25.3 |
| Released | 2024-02-21 | 2024-04-18 |
| Weights | Open | Open |
| Context window | — | 8K |
| Max output | — | 8K |
| Input $ / M tokens | — | $0.14 |
| Output $ / M tokens | — | $0.14 |
| Results tracked | 27 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 7B: 29.9 (#290), Llama 3-8B: 30.5 (#285)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Coding | 1048 | 1152 |
| HumanEval+ | 28.7% | 56.7% |
| MBPP+ | 43.4% | 54.8% |
| BigCodeBench Instruct | — | 31.9% |
| BigCodeBench Complete | — | 36.9% |
Reasoning Gemma 7B leads
Gemma 7B: 19.0 (#247), Llama 3-8B: 13.8 (#321)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Hard Prompts | 1042 | 1133 |
| Adversarial NLI | 48.7% | 57.3% |
| Epoch Capabilities Index | 111.74 | 116.33 |
| WinoGrande | 79% | 75.7% |
| Chess Puzzles | — | 0% |
| DTBench | — | 43.9% |
| BIG-Bench Hard | 55.1% | — |
| ForecastBench | — | 58.6 |
| HellaSwag | 82.2% | — |
| PIQA | 81.2% | — |
Math Gemma 7B leads
Gemma 7B: 28.7 (#225), Llama 3-8B: 8.3 (#317)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Math | 1066 | 1151 |
| OTIS Mock AIME 2024-2025 | — | 1.9% |
| MATH Level 5 | — | 6.1% |
| GSM8K | 46.4% | — |
Knowledge Gemma 7B leads
Gemma 7B: 26.8 (#248), Llama 3-8B: 7.8 (#303)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Expert | 1002 | 1113 |
| ARC (AI2) Challenge | 78.3% | 82.8% |
| MMLU | 66.1% | 68.8% |
| OpenBookQA | 78.6% | 82.6% |
| TriviaQA | 72.3% | 67.7% |
| GPQA Diamond | — | 26.1% |
| BoolQ | 83.2% | — |
Multilingual Llama 3-8B leads
Gemma 7B: 25.3 (#282), Llama 3-8B: 31.0 (#256)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 999 | 1098 |
| LMArena Chinese | 1035 | 1076 |
| LMArena French | 1024 | 1158 |
| LMArena Russian | 993 | 1109 |
| LMArena German | — | 1104 |
| LMArena Japanese | — | 967 |
| LMArena Korean | — | 1003 |
| LMArena Spanish | — | 1172 |
Instruction Following Llama 3-8B leads
Gemma 7B: 51.7 (#290), Llama 3-8B: 58.7 (#255)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1017 | 1127 |
Long Context Llama 3-8B leads
Gemma 7B: 31.1 (#277), Llama 3-8B: 34.2 (#246)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1022 | 1128 |
Writing & Preference Llama 3-8B leads
Gemma 7B: 27.2 (#297), Llama 3-8B: 37.7 (#251)
| Benchmark | Gemma 7B | Llama 3-8B |
|---|---|---|
| LMArena Text | 1056 | 1166 |
| LMArena Creative Writing | 1024 | 1150 |
| LMArena Multi-Turn | 963 | 1152 |
Frequently asked questions
Is Gemma 7B better than Llama 3-8B?
Gemma 7B is the stronger model overall, scoring 29.5 to 25.3 on the Noometry Index.
Is Gemma 7B or Llama 3-8B better for coding?
They score almost the same on coding (29.9 vs 30.5); test both on your own repository before choosing.
How many benchmarks do Gemma 7B and Llama 3-8B share?
22 benchmarks have published results for both models. Gemma 7B has 27 scored results on Noometry and Llama 3-8B has 34.