Model comparison
Mistral vs Mixtral 8x22B
Mistral is the stronger model overall, scoring 29.9 to 27.1 on the Noometry Index.
Last verified . 22 shared benchmarks.
Mistral AISummary
- They share 22 benchmarks with published results for both. Mistral scores higher in 6 categories and Mixtral 8x22B in 2 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in coding, where Mistral leads 33.8 to 24.2.
- The biggest single-benchmark swing is MMLU-Pro: 27.7% for Mistral and 46% for Mixtral 8x22B.
- Mixtral 8x22B has downloadable open weights; the other is API-only.
Side by side
| Mistral | Mixtral 8x22B | |
|---|---|---|
| Provider | Mistral AI | Mistral AI |
| Noometry Index | 29.9 | 27.1 |
| Released | — | 2024-04-17 |
| Weights | Proprietary | Open |
| Context window | — | 64K |
| Max output | — | 64K |
| Input $ / M tokens | — | $2 |
| Output $ / M tokens | — | $6 |
| Results tracked | 22 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral leads
Mistral: 33.8 (#250), Mixtral 8x22B: 24.2 (#329)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| LMArena Coding | 1162 | 1166 |
| WeirdML | — | 3.2% |
| BigCodeBench Instruct | — | 40.6% |
| BigCodeBench Complete | — | 50.2% |
| HumanEval+ | — | 72% |
| MBPP+ | — | 64.3% |
Agentic & Tool Use Not comparable
Mistral: —, Mixtral 8x22B: 23.1 (#127)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| Cybench | — | 7.5% |
Reasoning Mistral leads
Mistral: 22.2 (#200), Mixtral 8x22B: 19.9 (#248)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1149 | 1150 |
| DTBench | — | 55.1% |
| Epoch Capabilities Index | — | 122.03 |
| ForecastBench | — | 56.3 |
Math Too close to call
Mistral: 22.3 (#278), Mixtral 8x22B: 22.9 (#275)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| Omni-MATH | 7.2% | 16.3% |
| LMArena Math | 1180 | 1184 |
| MATH Level 5 | — | 24.2% |
Knowledge Mistral leads
Mistral: 16.6 (#288), Mixtral 8x22B: 15.1 (#293)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| MMLU-Pro | 27.7% | 46% |
| GPQA (HELM) | 30.3% | 33.4% |
| LMArena Expert | 1125 | 1113 |
| GPQA Diamond | — | 34.1% |
| MMLU | — | 77.8% |
Multilingual Too close to call
Mistral: 32.8 (#254), Mixtral 8x22B: 32.8 (#255)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1129 | 1128 |
| LMArena Chinese | 1109 | 1116 |
| LMArena French | 1180 | 1166 |
| LMArena German | 1155 | 1141 |
| LMArena Japanese | 1013 | 1037 |
| LMArena Korean | 1032 | 1057 |
| LMArena Russian | 1168 | 1158 |
| LMArena Spanish | 1143 | 1151 |
Instruction Following Mixtral 8x22B leads
Mistral: 52.6 (#288), Mixtral 8x22B: 57.7 (#266)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| IFEval | 56.8% | 72.4% |
| LMArena Instruction Following | 1152 | 1147 |
Long Context Too close to call
Mistral: 35.0 (#245), Mixtral 8x22B: 34.7 (#247)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1153 | 1144 |
Writing & Preference Too close to call
Mistral: 37.0 (#260), Mixtral 8x22B: 36.9 (#262)
| Benchmark | Mistral | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1165 | 1162 |
| LMArena Creative Writing | 1158 | 1141 |
| WildBench | 66% | 71.1% |
| LMArena Multi-Turn | 1147 | 1130 |
Frequently asked questions
Is Mistral better than Mixtral 8x22B?
Mistral is the stronger model overall, scoring 29.9 to 27.1 on the Noometry Index.
Is Mistral or Mixtral 8x22B better for coding?
Mistral scores higher on coding benchmarks: 33.8 versus 24.2 in the Noometry coding category.
How many benchmarks do Mistral and Mixtral 8x22B share?
22 benchmarks have published results for both models. Mistral has 22 scored results on Noometry and Mixtral 8x22B has 34.