Model comparison
Llama 3-8B vs Mixtral 8x7B
Mixtral 8x7B is the stronger model overall, scoring 27.1 to 25.5 on the Noometry Index.
Last verified . 30 shared benchmarks.
Meta
Mistral AISummary
- They share 30 benchmarks with published results for both. Llama 3-8B scores higher in 4 categories and Mixtral 8x7B in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Mixtral 8x7B leads 18.8 to 8.8.
- The biggest single-benchmark swing is DTBench: 43.9% for Llama 3-8B and 49.6% for Mixtral 8x7B.
Side by side
| Llama 3-8B | Mixtral 8x7B | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 25.5 | 27.1 |
| Released | 2024-04-18 | 2023-12-11 |
| Weights | Open | Open |
| Context window | — | 32K |
| Max output | — | 32K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $0.70 |
| Results tracked | 34 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mixtral 8x7B leads
Llama 3-8B: 31.0 (#289), Mixtral 8x7B: 32.8 (#269)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Coding | 1152 | 1126 |
| HumanEval+ | 56.7% | 39.6% |
| MBPP+ | 54.8% | 49.7% |
| BigCodeBench Instruct | 31.9% | — |
| BigCodeBench Complete | 36.9% | — |
Reasoning Mixtral 8x7B leads
Llama 3-8B: 14.3 (#326), Mixtral 8x7B: 18.2 (#285)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Hard Prompts | 1133 | 1115 |
| DTBench | 43.9% | 49.6% |
| Adversarial NLI | 57.3% | 55.2% |
| Epoch Capabilities Index | 116.45 | 118.47 |
| ForecastBench | 58.6 | 56.3 |
| WinoGrande | 75.7% | 77.2% |
| Chess Puzzles | 0% | — |
| HellaSwag | — | 86.7% |
| PIQA | — | 83.6% |
Math Mixtral 8x7B leads
Llama 3-8B: 8.8 (#323), Mixtral 8x7B: 18.8 (#289)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Math | 1151 | 1147 |
| MATH Level 5 | 6.1% | 10% |
| OTIS Mock AIME 2024-2025 | 1.9% | — |
| Omni-MATH | — | 10.5% |
| GSM8K | — | 74.4% |
Knowledge Mixtral 8x7B leads
Llama 3-8B: 7.8 (#308), Mixtral 8x7B: 11.0 (#301)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| GPQA Diamond | 26.1% | 30.6% |
| LMArena Expert | 1113 | 1088 |
| ARC (AI2) Challenge | 82.8% | 87.3% |
| MMLU | 68.8% | 70.6% |
| OpenBookQA | 82.6% | 85.8% |
| TriviaQA | 67.7% | 82.2% |
| MMLU-Pro | — | 33.5% |
| GPQA (HELM) | — | 29.6% |
Multilingual Llama 3-8B leads
Llama 3-8B: 30.8 (#261), Mixtral 8x7B: 29.6 (#266)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Non-English | 1098 | 1077 |
| LMArena Chinese | 1076 | 1055 |
| LMArena French | 1159 | 1166 |
| LMArena German | 1104 | 1114 |
| LMArena Japanese | 967 | 931 |
| LMArena Korean | 1004 | 968 |
| LMArena Russian | 1109 | 1090 |
| LMArena Spanish | 1173 | 1111 |
Instruction Following Llama 3-8B leads
Llama 3-8B: 58.4 (#260), Mixtral 8x7B: 51.0 (#297)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Instruction Following | 1127 | 1109 |
| IFEval | — | 57.5% |
Long Context Too close to call
Llama 3-8B: 34.2 (#251), Mixtral 8x7B: 33.4 (#260)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Longer Query | 1128 | 1103 |
Writing & Preference Llama 3-8B leads
Llama 3-8B: 37.5 (#256), Mixtral 8x7B: 34.2 (#270)
| Benchmark | Llama 3-8B | Mixtral 8x7B |
|---|---|---|
| LMArena Text | 1166 | 1132 |
| LMArena Creative Writing | 1150 | 1109 |
| LMArena Multi-Turn | 1152 | 1115 |
| WildBench | — | 67.3% |
Frequently asked questions
Is Llama 3-8B better than Mixtral 8x7B?
Mixtral 8x7B is the stronger model overall, scoring 27.1 to 25.5 on the Noometry Index.
Is Llama 3-8B or Mixtral 8x7B better for coding?
Mixtral 8x7B scores higher on coding benchmarks: 32.8 versus 31.0 in the Noometry coding category.
How many benchmarks do Llama 3-8B and Mixtral 8x7B share?
30 benchmarks have published results for both models. Llama 3-8B has 34 scored results on Noometry and Mixtral 8x7B has 38.