Model comparison
Llama 3-70B vs Mixtral 8x22B
Llama 3-70B is the stronger model overall, scoring 28.8 to 27.1 on the Noometry Index.
Last verified . 28 shared benchmarks.
Meta
Mistral AISummary
- They share 28 benchmarks with published results for both. Llama 3-70B scores higher in 6 categories and Mixtral 8x22B in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Llama 3-70B leads 35.8 to 24.2.
- The biggest single-benchmark swing is GPQA Diamond: 40.6% for Llama 3-70B and 34.1% for Mixtral 8x22B.
Side by side
| Llama 3-70B | Mixtral 8x22B | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 28.8 | 27.1 |
| Released | 2024-04-18 | 2024-04-17 |
| Weights | Open | Open |
| Context window | — | 64K |
| Max output | — | 64K |
| Input $ / M tokens | — | $2 |
| Output $ / M tokens | — | $6 |
| Results tracked | 31 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3-70B leads
Llama 3-70B: 35.8 (#218), Mixtral 8x22B: 24.2 (#329)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| BigCodeBench Instruct | 43.6% | 40.6% |
| LMArena Coding | 1206 | 1166 |
| BigCodeBench Complete | 54.5% | 50.2% |
| HumanEval+ | 72% | 72% |
| MBPP+ | 69% | 64.3% |
| WeirdML | — | 3.2% |
Agentic & Tool Use Mixtral 8x22B leads
Llama 3-70B: 21.1 (#139), Mixtral 8x22B: 23.1 (#127)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| Cybench | 5% | 7.5% |
Reasoning Mixtral 8x22B leads
Llama 3-70B: 18.0 (#288), Mixtral 8x22B: 19.9 (#248)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1195 | 1150 |
| DTBench | 54.2% | 55.1% |
| Epoch Capabilities Index | 122.93 | 122.03 |
| ForecastBench | 57.1 | 56.3 |
| Kagi LLM Benchmark | 35.1% | — |
| WinoGrande | 83.5% | — |
Math Mixtral 8x22B leads
Llama 3-70B: 12.8 (#305), Mixtral 8x22B: 22.9 (#275)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Math | 1218 | 1184 |
| MATH Level 5 | 22.6% | 24.2% |
| OTIS Mock AIME 2024-2025 | 4.3% | — |
| Omni-MATH | — | 16.3% |
Knowledge Llama 3-70B leads
Llama 3-70B: 20.8 (#277), Mixtral 8x22B: 15.1 (#293)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 40.6% | 34.1% |
| LMArena Expert | 1149 | 1113 |
| MMLU | 79.3% | 77.8% |
| MMLU-Pro | — | 46% |
| GPQA (HELM) | — | 33.4% |
Multilingual Too close to call
Llama 3-70B: 33.6 (#251), Mixtral 8x22B: 32.8 (#255)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1142 | 1128 |
| LMArena Chinese | 1114 | 1116 |
| LMArena French | 1232 | 1166 |
| LMArena German | 1169 | 1141 |
| LMArena Japanese | 1017 | 1037 |
| LMArena Korean | 1017 | 1057 |
| LMArena Russian | 1159 | 1158 |
| LMArena Spanish | 1241 | 1151 |
Instruction Following Llama 3-70B leads
Llama 3-70B: 62.5 (#238), Mixtral 8x22B: 57.7 (#266)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Instruction Following | 1194 | 1147 |
| IFEval | — | 72.4% |
Long Context Too close to call
Llama 3-70B: 35.6 (#240), Mixtral 8x22B: 34.7 (#247)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1174 | 1144 |
Writing & Preference Llama 3-70B leads
Llama 3-70B: 42.8 (#231), Mixtral 8x22B: 36.9 (#262)
| Benchmark | Llama 3-70B | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1221 | 1162 |
| LMArena Creative Writing | 1210 | 1141 |
| LMArena Multi-Turn | 1223 | 1130 |
| WildBench | — | 71.1% |
Frequently asked questions
Is Llama 3-70B better than Mixtral 8x22B?
Llama 3-70B is the stronger model overall, scoring 28.8 to 27.1 on the Noometry Index.
Is Llama 3-70B or Mixtral 8x22B better for coding?
Llama 3-70B scores higher on coding benchmarks: 35.8 versus 24.2 in the Noometry coding category.
How many benchmarks do Llama 3-70B and Mixtral 8x22B share?
28 benchmarks have published results for both models. Llama 3-70B has 31 scored results on Noometry and Mixtral 8x22B has 34.