Model comparison
Llama 2-70B vs Mixtral 8x22B
Mixtral 8x22B is the stronger model overall, scoring 27.1 to 24.4 on the Noometry Index.
Last verified . 23 shared benchmarks.
Meta
Mistral AISummary
- They share 23 benchmarks with published results for both. Llama 2-70B scores higher in 1 category and Mixtral 8x22B in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Mixtral 8x22B leads 22.9 to 8.1.
- The biggest single-benchmark swing is MATH Level 5: 3.3% for Llama 2-70B and 24.2% for Mixtral 8x22B.
Side by side
| Llama 2-70B | Mixtral 8x22B | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 24.4 | 27.1 |
| Released | 2023-07-18 | 2024-04-17 |
| Weights | Open | Open |
| Context window | — | 64K |
| Max output | — | 64K |
| Input $ / M tokens | — | $2 |
| Output $ / M tokens | — | $6 |
| Results tracked | 35 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-70B leads
Llama 2-70B: 31.4 (#286), Mixtral 8x22B: 24.2 (#329)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Coding | 1079 | 1166 |
| WeirdML | — | 3.2% |
| BigCodeBench Instruct | — | 40.6% |
| BigCodeBench Complete | — | 50.2% |
| HumanEval+ | — | 72% |
| MBPP+ | — | 64.3% |
Agentic & Tool Use Not comparable
Llama 2-70B: —, Mixtral 8x22B: 23.1 (#127)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| Cybench | — | 7.5% |
Reasoning Mixtral 8x22B leads
Llama 2-70B: 14.4 (#325), Mixtral 8x22B: 19.9 (#248)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1150 |
| DTBench | 41.6% | 55.1% |
| Epoch Capabilities Index | 113.79 | 122.03 |
| ForecastBench | 51.4 | 56.3 |
| BIG-Bench Hard | 64.9% | — |
| CommonsenseQA 2.0 | 50% | — |
| HellaSwag | 85.3% | — |
| LAMBADA | 78.9% | — |
| PIQA | 82.8% | — |
| WinoGrande | 80.2% | — |
Math Mixtral 8x22B leads
Llama 2-70B: 8.1 (#326), Mixtral 8x22B: 22.9 (#275)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Math | 1091 | 1184 |
| MATH Level 5 | 3.3% | 24.2% |
| OTIS Mock AIME 2024-2025 | 0% | — |
| Omni-MATH | — | 16.3% |
| GSM8K | 69.6% | — |
Knowledge Mixtral 8x22B leads
Llama 2-70B: 7.4 (#310), Mixtral 8x22B: 15.1 (#293)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 26.3% | 34.1% |
| LMArena Expert | 1039 | 1113 |
| MMLU | 69.9% | 77.8% |
| MMLU-Pro | — | 46% |
| GPQA (HELM) | — | 33.4% |
| ARC (AI2) Challenge | 78.3% | — |
| BoolQ | 88.6% | — |
| OpenBookQA | 60.2% | — |
| TriviaQA | 87.6% | — |
Multilingual Mixtral 8x22B leads
Llama 2-70B: 27.7 (#274), Mixtral 8x22B: 32.8 (#255)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1045 | 1128 |
| LMArena Chinese | 995 | 1116 |
| LMArena French | 1090 | 1166 |
| LMArena German | 1041 | 1141 |
| LMArena Japanese | 927 | 1037 |
| LMArena Korean | 964 | 1057 |
| LMArena Russian | 1083 | 1158 |
| LMArena Spanish | 1143 | 1151 |
Instruction Following Mixtral 8x22B leads
Llama 2-70B: 54.9 (#278), Mixtral 8x22B: 57.7 (#266)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1147 |
| IFEval | — | 72.4% |
Long Context Mixtral 8x22B leads
Llama 2-70B: 32.3 (#270), Mixtral 8x22B: 34.7 (#247)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1062 | 1144 |
Writing & Preference Mixtral 8x22B leads
Llama 2-70B: 32.3 (#279), Mixtral 8x22B: 36.9 (#262)
| Benchmark | Llama 2-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1115 | 1162 |
| LMArena Creative Writing | 1075 | 1141 |
| LMArena Multi-Turn | 1088 | 1130 |
| WildBench | — | 71.1% |
Frequently asked questions
Is Llama 2-70B better than Mixtral 8x22B?
Mixtral 8x22B is the stronger model overall, scoring 27.1 to 24.4 on the Noometry Index.
Is Llama 2-70B or Mixtral 8x22B better for coding?
Llama 2-70B scores higher on coding benchmarks: 31.4 versus 24.2 in the Noometry coding category.
How many benchmarks do Llama 2-70B and Mixtral 8x22B share?
23 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and Mixtral 8x22B has 34.