Model comparison
Gemma 2 9B vs Llama 3.1-70B
Llama 3.1-70B is the stronger model overall, scoring 29.4 to 25.7 on the Noometry Index.
Last verified . 25 shared benchmarks.
Google
MetaSummary
- They share 25 benchmarks with published results for both. Gemma 2 9B scores higher in 0 categories and Llama 3.1-70B in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 3.1-70B leads 24.1 to 9.7.
- The biggest single-benchmark swing is GPQA Diamond: 27.5% for Gemma 2 9B and 44.2% for Llama 3.1-70B.
Side by side
| Gemma 2 9B | Llama 3.1-70B | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 25.7 | 29.4 |
| Released | 2024-06-24 | 2024-07-23 |
| Weights | Open | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.40 |
| Output $ / M tokens | — | $0.40 |
| Results tracked | 35 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 2 9B: 29.0 (#299), Llama 3.1-70B: 29.8 (#292)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| BigCodeBench Instruct | 34.7% | 46.1% |
| LMArena Coding | 1174 | 1260 |
| BigCodeBench Complete | 40.6% | 54.8% |
| WeirdML | — | 9% |
| LiveBench Coding | 22.5% | — |
Agentic & Tool Use Not comparable
Gemma 2 9B: —, Llama 3.1-70B: 24.9 (#104)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| TheAgentCompany | — | 6.9% |
| BALROG | — | 27.9% |
Reasoning Llama 3.1-70B leads
Gemma 2 9B: 15.3 (#306), Llama 3.1-70B: 20.8 (#215)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| LMArena Hard Prompts | 1172 | 1241 |
| Epoch Capabilities Index | 119.78 | 125.91 |
| LiveBench Reasoning | 15.2% | — |
| DTBench | — | 60% |
| LiveBench Data Analysis | 36.4% | — |
| LMCA | — | 14.8% |
| LiveBench | 28.7% | — |
| PIQA | 83.7% | — |
Math Llama 3.1-70B leads
Gemma 2 9B: 9.3 (#312), Llama 3.1-70B: 12.7 (#298)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.6% | 3.6% |
| LMArena Math | 1183 | 1252 |
| MATH Level 5 | 21% | 36.7% |
| Omni-MATH | — | 21% |
| LiveBench Math | 19.8% | — |
| GSM8K | 84.9% | — |
Knowledge Llama 3.1-70B leads
Gemma 2 9B: 9.7 (#300), Llama 3.1-70B: 24.1 (#264)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| GPQA Diamond | 27.5% | 44.2% |
| LMArena Expert | 1147 | 1209 |
| MMLU | 72.1% | 80.1% |
| MMLU-Pro | — | 65.3% |
| GPQA (HELM) | — | 42.6% |
| BoolQ | 85.7% | — |
Multilingual Llama 3.1-70B leads
Gemma 2 9B: 36.8 (#233), Llama 3.1-70B: 38.9 (#220)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| LMArena Non-English | 1188 | 1219 |
| LMArena Chinese | 1185 | 1215 |
| LMArena French | 1189 | 1260 |
| LMArena German | 1186 | 1222 |
| LMArena Japanese | 1144 | 1132 |
| LMArena Korean | 1137 | 1140 |
| LMArena Russian | 1200 | 1234 |
| LMArena Spanish | 1200 | 1253 |
Instruction Following Llama 3.1-70B leads
Gemma 2 9B: 57.8 (#264), Llama 3.1-70B: 65.5 (#218)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| LMArena Instruction Following | 1178 | 1231 |
| LiveBench Instruction Following | 52.6% | — |
| IFEval | — | 82.1% |
Long Context Llama 3.1-70B leads
Gemma 2 9B: 36.3 (#228), Llama 3.1-70B: 37.6 (#208)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| LMArena Longer Query | 1197 | 1241 |
Writing & Preference Llama 3.1-70B leads
Gemma 2 9B: 32.3 (#276), Llama 3.1-70B: 35.6 (#262)
| Benchmark | Gemma 2 9B | Llama 3.1-70B |
|---|---|---|
| LMArena Text | 1207 | 1261 |
| LMArena Creative Writing | 1206 | 1232 |
| EQ-Bench Creative Writing | 841 | 784 |
| LMArena Multi-Turn | 1193 | 1256 |
| WildBench | — | 75.8% |
| LiveBench Language | 25.5% | — |
Frequently asked questions
Is Gemma 2 9B better than Llama 3.1-70B?
Llama 3.1-70B is the stronger model overall, scoring 29.4 to 25.7 on the Noometry Index.
Is Gemma 2 9B or Llama 3.1-70B better for coding?
They score almost the same on coding (29.0 vs 29.8); test both on your own repository before choosing.
How many benchmarks do Gemma 2 9B and Llama 3.1-70B share?
25 benchmarks have published results for both models. Gemma 2 9B has 35 scored results on Noometry and Llama 3.1-70B has 35.