Model comparison
Mistral Large vs Mixtral 8x22B
Mistral Large is the stronger model overall, scoring 31.9 to 27.1 on the Noometry Index.
Last verified . 32 shared benchmarks.
Mistral AISummary
- They share 32 benchmarks with published results for both. Mistral Large scores higher in 7 categories and Mixtral 8x22B in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Mistral Large leads 30.1 to 15.1.
- The biggest single-benchmark swing is MATH Level 5: 50.3% for Mistral Large and 24.2% for Mixtral 8x22B.
- Both cost about the same: $2 input and $6 output per million tokens.
- Mistral Large accepts more context: 131K tokens versus 64K.
Side by side
| Mistral Large | Mixtral 8x22B | |
|---|---|---|
| Provider | Mistral AI | Mistral AI |
| Noometry Index | 31.9 | 27.1 |
| Released | 2024-02-26 | 2024-04-17 |
| Weights | Open | Open |
| Context window | 131K | 64K |
| Max output | 16K | 64K |
| Input $ / M tokens | $2 | $2 |
| Output $ / M tokens | $6 | $6 |
| Results tracked | 51 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Large leads
Mistral Large: 34.3 (#240), Mixtral 8x22B: 24.2 (#329)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| BigCodeBench Instruct | 30% | 40.6% |
| LMArena Coding | 1277 | 1166 |
| BigCodeBench Complete | 38.3% | 50.2% |
| HumanEval+ | 62.2% | 72% |
| MBPP+ | 59.5% | 64.3% |
| SciCode | 36.2% | — |
| WeirdML | — | 3.2% |
| LiveBench Coding | 47.1% | — |
| ALE-Bench | 264.7 | — |
Agentic & Tool Use Mistral Large leads
Mistral Large: 28.6 (#89), Mixtral 8x22B: 23.1 (#127)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 38.4% | — |
| Cybench | — | 7.5% |
Reasoning Mixtral 8x22B leads
Mistral Large: 15.8 (#310), Mixtral 8x22B: 19.9 (#248)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1257 | 1150 |
| DTBench | 65.1% | 55.1% |
| Epoch Capabilities Index | 128.52 | 122.03 |
| ForecastBench | 57.1 | 56.3 |
| SimpleBench | 22.5% | — |
| CritPt | 0% | — |
| LiveBench Reasoning | 43.5% | — |
| LiveBench Data Analysis | 50.1% | — |
| LMCA | 16.7% | — |
| LiveBench | 48.4% | — |
Math Mixtral 8x22B leads
Mistral Large: 18.2 (#291), Mixtral 8x22B: 22.9 (#275)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| Omni-MATH | 28.1% | 16.3% |
| LMArena Math | 1262 | 1184 |
| MATH Level 5 | 50.3% | 24.2% |
| OTIS Mock AIME 2024-2025 | 8.5% | — |
| LiveBench Math | 42.5% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
Knowledge Mistral Large leads
Mistral Large: 30.1 (#230), Mixtral 8x22B: 15.1 (#293)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 51.3% | 34.1% |
| MMLU-Pro | 59.9% | 46% |
| GPQA (HELM) | 43.5% | 33.4% |
| LMArena Expert | 1232 | 1113 |
| MMLU | 80% | 77.8% |
| Confabulations | 21.4% | — |
| Vectara Hallucination Rate | 4.5% | — |
Multilingual Mistral Large leads
Mistral Large: 40.0 (#219), Mixtral 8x22B: 32.8 (#255)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1237 | 1128 |
| LMArena Chinese | 1240 | 1116 |
| LMArena French | 1325 | 1166 |
| LMArena German | 1254 | 1141 |
| LMArena Japanese | 1188 | 1037 |
| LMArena Korean | 1202 | 1057 |
| LMArena Russian | 1257 | 1158 |
| LMArena Spanish | 1268 | 1151 |
Instruction Following Mistral Large leads
Mistral Large: 67.9 (#191), Mixtral 8x22B: 57.7 (#266)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| IFEval | 87.7% | 72.4% |
| LMArena Instruction Following | 1249 | 1147 |
| LiveBench Instruction Following | 67.9% | — |
Long Context Mistral Large leads
Mistral Large: 38.3 (#199), Mixtral 8x22B: 34.7 (#247)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1261 | 1144 |
Writing & Preference Mistral Large leads
Mistral Large: 40.7 (#242), Mixtral 8x22B: 36.9 (#262)
| Benchmark | Mistral Large | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1266 | 1162 |
| LMArena Creative Writing | 1243 | 1141 |
| WildBench | 80.1% | 71.1% |
| LMArena Multi-Turn | 1260 | 1130 |
| Short-Story Creative Writing | 69% | — |
| EQ-Bench Creative Writing | 985 | — |
| LiveBench Language | 39.4% | — |
Frequently asked questions
Is Mistral Large better than Mixtral 8x22B?
Mistral Large is the stronger model overall, scoring 31.9 to 27.1 on the Noometry Index.
Which is cheaper, Mistral Large or Mixtral 8x22B?
Mixtral 8x22B is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; Mistral Large lists at $2 and $6.
Is Mistral Large or Mixtral 8x22B better for coding?
Mistral Large scores higher on coding benchmarks: 34.3 versus 24.2 in the Noometry coding category.
Which has the bigger context window?
Mistral Large does, with 131K tokens against 64K.
How many benchmarks do Mistral Large and Mixtral 8x22B share?
32 benchmarks have published results for both models. Mistral Large has 51 scored results on Noometry and Mixtral 8x22B has 34.