Model comparison
Gemma 2 9B vs Mixtral 8x7B
Mixtral 8x7B is the stronger model overall, scoring 27.1 to 25.9 on the Noometry Index.
Last verified . 23 shared benchmarks.
Google
Mistral AISummary
- They share 23 benchmarks with published results for both. Gemma 2 9B scores higher in 3 categories and Mixtral 8x7B in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Mixtral 8x7B leads 18.8 to 9.9.
- The biggest single-benchmark swing is MATH Level 5: 21% for Gemma 2 9B and 10% for Mixtral 8x7B.
Side by side
| Gemma 2 9B | Mixtral 8x7B | |
|---|---|---|
| Provider | Mistral AI | |
| Noometry Index | 25.9 | 27.1 |
| Released | 2024-06-24 | 2023-12-11 |
| Weights | Open | Open |
| Context window | — | 32K |
| Max output | — | 32K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $0.70 |
| Results tracked | 35 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mixtral 8x7B leads
Gemma 2 9B: 29.4 (#304), Mixtral 8x7B: 32.8 (#269)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Coding | 1173 | 1126 |
| BigCodeBench Instruct | 34.7% | — |
| LiveBench Coding | 22.5% | — |
| BigCodeBench Complete | 40.6% | — |
| HumanEval+ | — | 39.6% |
| MBPP+ | — | 49.7% |
Reasoning Mixtral 8x7B leads
Gemma 2 9B: 15.9 (#309), Mixtral 8x7B: 18.2 (#285)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Hard Prompts | 1171 | 1115 |
| Epoch Capabilities Index | 119.83 | 118.47 |
| PIQA | 83.7% | 83.6% |
| LiveBench Reasoning | 15.2% | — |
| DTBench | — | 49.6% |
| LiveBench Data Analysis | 36.4% | — |
| Adversarial NLI | — | 55.2% |
| ForecastBench | — | 56.3 |
| HellaSwag | — | 86.7% |
| LiveBench | 28.7% | — |
| WinoGrande | — | 77.2% |
Math Mixtral 8x7B leads
Gemma 2 9B: 9.9 (#318), Mixtral 8x7B: 18.8 (#289)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Math | 1183 | 1147 |
| MATH Level 5 | 21% | 10% |
| GSM8K | 84.9% | 74.4% |
| OTIS Mock AIME 2024-2025 | 0.6% | — |
| Omni-MATH | — | 10.5% |
| LiveBench Math | 19.8% | — |
Knowledge Mixtral 8x7B leads
Gemma 2 9B: 9.7 (#305), Mixtral 8x7B: 11.0 (#301)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| GPQA Diamond | 27.5% | 30.6% |
| LMArena Expert | 1147 | 1088 |
| MMLU | 72.1% | 70.6% |
| MMLU-Pro | — | 33.5% |
| GPQA (HELM) | — | 29.6% |
| ARC (AI2) Challenge | — | 87.3% |
| BoolQ | 85.7% | — |
| OpenBookQA | — | 85.8% |
| TriviaQA | — | 82.2% |
Multilingual Gemma 2 9B leads
Gemma 2 9B: 36.6 (#238), Mixtral 8x7B: 29.6 (#266)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Non-English | 1188 | 1077 |
| LMArena Chinese | 1185 | 1055 |
| LMArena French | 1190 | 1166 |
| LMArena German | 1186 | 1114 |
| LMArena Japanese | 1144 | 931 |
| LMArena Korean | 1137 | 968 |
| LMArena Russian | 1200 | 1090 |
| LMArena Spanish | 1200 | 1111 |
Instruction Following Gemma 2 9B leads
Gemma 2 9B: 57.6 (#269), Mixtral 8x7B: 51.0 (#297)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Instruction Following | 1178 | 1109 |
| LiveBench Instruction Following | 52.6% | — |
| IFEval | — | 57.5% |
Long Context Gemma 2 9B leads
Gemma 2 9B: 36.3 (#233), Mixtral 8x7B: 33.4 (#260)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Longer Query | 1197 | 1103 |
Writing & Preference Mixtral 8x7B leads
Gemma 2 9B: 32.1 (#281), Mixtral 8x7B: 34.2 (#270)
| Benchmark | Gemma 2 9B | Mixtral 8x7B |
|---|---|---|
| LMArena Text | 1207 | 1132 |
| LMArena Creative Writing | 1206 | 1109 |
| LMArena Multi-Turn | 1193 | 1115 |
| EQ-Bench Creative Writing | 841 | — |
| WildBench | — | 67.3% |
| LiveBench Language | 25.5% | — |
Frequently asked questions
Is Gemma 2 9B better than Mixtral 8x7B?
Mixtral 8x7B is the stronger model overall, scoring 27.1 to 25.9 on the Noometry Index.
Is Gemma 2 9B or Mixtral 8x7B better for coding?
Mixtral 8x7B scores higher on coding benchmarks: 32.8 versus 29.4 in the Noometry coding category.
How many benchmarks do Gemma 2 9B and Mixtral 8x7B share?
23 benchmarks have published results for both models. Gemma 2 9B has 35 scored results on Noometry and Mixtral 8x7B has 38.