Model comparison
Mistral Large vs Mixtral 8x7B
Mistral Large is the stronger model overall, scoring 31.6 to 26.8 on the Noometry Index. Mixtral 8x7B costs 4.3× less per token, which makes it the better buy when Mistral Large's lead doesn't matter for your workload.
Last verified . 30 shared benchmarks.
Mistral AISummary
- They share 30 benchmarks with published results for both. Mistral Large scores higher in 6 categories and Mixtral 8x7B in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Mistral Large leads 30.0 to 11.0.
- The biggest single-benchmark swing is MATH Level 5: 50.3% for Mistral Large and 10% for Mixtral 8x7B.
- Mixtral 8x7B is cheaper at $0.70 / $0.70 per million input/output tokens, against $2 / $6 for Mistral Large.
- Mistral Large accepts more context: 131K tokens versus 32K.
Side by side
| Mistral Large | Mixtral 8x7B | |
|---|---|---|
| Provider | Mistral AI | Mistral AI |
| Noometry Index | 31.6 | 26.8 |
| Released | 2024-02-26 | 2023-12-11 |
| Weights | Open | Open |
| Context window | 131K | 32K |
| Max output | 16K | 32K |
| Input $ / M tokens | $2 | $0.70 |
| Output $ / M tokens | $6 | $0.70 |
| Results tracked | 51 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Large leads
Mistral Large: 33.8 (#240), Mixtral 8x7B: 32.2 (#265)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| LMArena Coding | 1277 | 1127 |
| HumanEval+ | 62.2% | 39.6% |
| MBPP+ | 59.5% | 49.7% |
| SciCode | 36.2% | — |
| BigCodeBench Instruct | 30% | — |
| LiveBench Coding | 47.1% | — |
| BigCodeBench Complete | 38.3% | — |
| ALE-Bench | 264.7 | — |
Agentic & Tool Use Not comparable
Mistral Large: 28.4 (#83), Mixtral 8x7B: —
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 38.4% | — |
Reasoning Mixtral 8x7B leads
Mistral Large: 15.4 (#305), Mixtral 8x7B: 17.4 (#283)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| LMArena Hard Prompts | 1257 | 1115 |
| DTBench | 65.1% | 49.6% |
| Epoch Capabilities Index | 128.52 | 118.38 |
| ForecastBench | 57.1 | 56.3 |
| SimpleBench | 22.5% | — |
| CritPt | 0% | — |
| LiveBench Reasoning | 43.5% | — |
| LiveBench Data Analysis | 50.1% | — |
| LMCA | 16.7% | — |
| Adversarial NLI | — | 55.2% |
| HellaSwag | — | 86.7% |
| LiveBench | 48.4% | — |
| PIQA | — | 83.6% |
| WinoGrande | — | 77.2% |
Math Too close to call
Mistral Large: 17.1 (#285), Mixtral 8x7B: 17.2 (#284)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| Omni-MATH | 28.1% | 10.5% |
| LMArena Math | 1262 | 1147 |
| MATH Level 5 | 50.3% | 10% |
| OTIS Mock AIME 2024-2025 | 8.5% | — |
| LiveBench Math | 42.5% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
| GSM8K | — | 74.4% |
Knowledge Mistral Large leads
Mistral Large: 30.0 (#224), Mixtral 8x7B: 11.0 (#296)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| GPQA Diamond | 51.3% | 30.6% |
| MMLU-Pro | 59.9% | 33.5% |
| GPQA (HELM) | 43.5% | 29.6% |
| LMArena Expert | 1232 | 1088 |
| MMLU | 80% | 70.6% |
| Confabulations | 21.4% | — |
| Vectara Hallucination Rate | 4.5% | — |
| ARC (AI2) Challenge | — | 87.3% |
| OpenBookQA | — | 85.8% |
| TriviaQA | — | 82.2% |
Multilingual Mistral Large leads
Mistral Large: 40.1 (#214), Mixtral 8x7B: 29.8 (#261)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| LMArena Non-English | 1237 | 1077 |
| LMArena Chinese | 1240 | 1055 |
| LMArena French | 1324 | 1165 |
| LMArena German | 1254 | 1114 |
| LMArena Japanese | 1187 | 931 |
| LMArena Korean | 1202 | 968 |
| LMArena Russian | 1257 | 1090 |
| LMArena Spanish | 1268 | 1110 |
Instruction Following Mistral Large leads
Mistral Large: 68.1 (#186), Mixtral 8x7B: 51.2 (#292)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| IFEval | 87.7% | 57.5% |
| LMArena Instruction Following | 1249 | 1110 |
| LiveBench Instruction Following | 67.9% | — |
Long Context Mistral Large leads
Mistral Large: 38.3 (#194), Mixtral 8x7B: 33.4 (#255)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| LMArena Longer Query | 1262 | 1103 |
Writing & Preference Mistral Large leads
Mistral Large: 40.9 (#237), Mixtral 8x7B: 34.4 (#265)
| Benchmark | Mistral Large | Mixtral 8x7B |
|---|---|---|
| LMArena Text | 1266 | 1132 |
| LMArena Creative Writing | 1243 | 1109 |
| WildBench | 80.1% | 67.3% |
| LMArena Multi-Turn | 1260 | 1115 |
| Short-Story Creative Writing | 69% | — |
| EQ-Bench Creative Writing | 985 | — |
| LiveBench Language | 39.4% | — |
Frequently asked questions
Is Mistral Large better than Mixtral 8x7B?
Mistral Large is the stronger model overall, scoring 31.6 to 26.8 on the Noometry Index. Mixtral 8x7B costs 4.3× less per token, which makes it the better buy when Mistral Large's lead doesn't matter for your workload.
Which is cheaper, Mistral Large or Mixtral 8x7B?
Mixtral 8x7B is cheaper. It lists at $0.70 per million input tokens and $0.70 per million output tokens; Mistral Large lists at $2 and $6.
Is Mistral Large or Mixtral 8x7B better for coding?
Mistral Large scores higher on coding benchmarks: 33.8 versus 32.2 in the Noometry coding category.
Which has the bigger context window?
Mistral Large does, with 131K tokens against 32K.
How many benchmarks do Mistral Large and Mixtral 8x7B share?
30 benchmarks have published results for both models. Mistral Large has 51 scored results on Noometry and Mixtral 8x7B has 38.