Model comparison
Mistral Small vs Mixtral 8x22B
Mistral Small is the stronger model overall, scoring 33.4 to 27.1 on the Noometry Index.
Last verified . 23 shared benchmarks.
Mistral AISummary
- They share 23 benchmarks with published results for both. Mistral Small scores higher in 7 categories and Mixtral 8x22B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Mistral Small leads 31.0 to 15.1.
- The biggest single-benchmark swing is MATH Level 5: 46.8% for Mistral Small and 24.2% for Mixtral 8x22B.
- Mistral Small is cheaper at $0.15 / $0.60 per million input/output tokens, against $2 / $6 for Mixtral 8x22B.
- Mistral Small accepts more context: 262K tokens versus 64K.
Side by side
| Mistral Small | Mixtral 8x22B | |
|---|---|---|
| Provider | Mistral AI | Mistral AI |
| Noometry Index | 33.4 | 27.1 |
| Released | 2024-02-26 | 2024-04-17 |
| Weights | Open | Open |
| Context window | 262K | 64K |
| Max output | 256K | 64K |
| Input $ / M tokens | $0.15 | $2 |
| Output $ / M tokens | $0.60 | $6 |
| Results tracked | 39 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Small leads
Mistral Small: 34.0 (#247), Mixtral 8x22B: 24.2 (#329)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| BigCodeBench Instruct | 36.1% | 40.6% |
| LMArena Coding | 1362 | 1166 |
| BigCodeBench Complete | 46.6% | 50.2% |
| SciCode | 26.5% | — |
| WeirdML | — | 3.2% |
| LiveBench Coding | 36.2% | — |
| ALE-Bench | 497.62 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 64.3% |
Agentic & Tool Use Mistral Small leads
Mistral Small: 28.1 (#93), Mixtral 8x22B: 23.1 (#127)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 37.1% | — |
| Cybench | — | 7.5% |
Reasoning Too close to call
Mistral Small: 19.8 (#250), Mixtral 8x22B: 19.9 (#248)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1335 | 1150 |
| DTBench | 70.9% | 55.1% |
| Kagi LLM Benchmark | 37.8% | — |
| CritPt | 0% | — |
| LiveBench Reasoning | 44.8% | — |
| LiveBench Data Analysis | 53.7% | — |
| LMCA | 20.6% | — |
| Epoch Capabilities Index | — | 122.03 |
| ForecastBench | — | 56.3 |
| LiveBench | 44% | — |
Math Mixtral 8x22B leads
Mistral Small: 16.4 (#293), Mixtral 8x22B: 22.9 (#275)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Math | 1341 | 1184 |
| MATH Level 5 | 46.8% | 24.2% |
| OTIS Mock AIME 2024-2025 | 5.8% | — |
| Omni-MATH | — | 16.3% |
| LiveBench Math | 39.9% | — |
Knowledge Mistral Small leads
Mistral Small: 31.0 (#222), Mixtral 8x22B: 15.1 (#293)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 47.5% | 34.1% |
| LMArena Expert | 1291 | 1113 |
| MMLU | 68.7% | 77.8% |
| MMLU-Pro | — | 46% |
| Vectara Hallucination Rate | 5.1% | — |
| GPQA (HELM) | — | 33.4% |
Multimodal Not comparable
Mistral Small: 33.5 (#96), Mixtral 8x22B: —
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Vision | 1142 | — |
Multilingual Mistral Small leads
Mistral Small: 45.5 (#169), Mixtral 8x22B: 32.8 (#255)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1315 | 1128 |
| LMArena Chinese | 1340 | 1116 |
| LMArena French | 1337 | 1166 |
| LMArena German | 1340 | 1141 |
| LMArena Japanese | 1275 | 1037 |
| LMArena Korean | 1259 | 1057 |
| LMArena Russian | 1324 | 1158 |
| LMArena Spanish | 1346 | 1151 |
Instruction Following Mistral Small leads
Mistral Small: 66.4 (#209), Mixtral 8x22B: 57.7 (#266)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Instruction Following | 1310 | 1147 |
| LiveBench Instruction Following | 63.7% | — |
| IFEval | — | 72.4% |
Long Context Mistral Small leads
Mistral Small: 40.4 (#156), Mixtral 8x22B: 34.7 (#247)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1327 | 1144 |
Writing & Preference Mistral Small leads
Mistral Small: 52.5 (#171), Mixtral 8x22B: 36.9 (#262)
| Benchmark | Mistral Small | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1338 | 1162 |
| LMArena Creative Writing | 1305 | 1141 |
| LMArena Multi-Turn | 1344 | 1130 |
| WildBench | — | 71.1% |
| LiveBench Language | 30.5% | — |
Frequently asked questions
Is Mistral Small better than Mixtral 8x22B?
Mistral Small is the stronger model overall, scoring 33.4 to 27.1 on the Noometry Index.
Which is cheaper, Mistral Small or Mixtral 8x22B?
Mistral Small is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Mixtral 8x22B lists at $2 and $6.
Is Mistral Small or Mixtral 8x22B better for coding?
Mistral Small scores higher on coding benchmarks: 34.0 versus 24.2 in the Noometry coding category.
Which has the bigger context window?
Mistral Small does, with 262K tokens against 64K.
How many benchmarks do Mistral Small and Mixtral 8x22B share?
23 benchmarks have published results for both models. Mistral Small has 39 scored results on Noometry and Mixtral 8x22B has 34.