Model comparison
Mixtral 8x22B vs Qwen1.5-32B
Qwen1.5-32B is the stronger model overall, scoring 30.0 to 26.7 on the Noometry Index.
Last verified . 21 shared benchmarks.
Mistral AI
Alibaba (Qwen)Summary
- They share 21 benchmarks with published results for both. Mixtral 8x22B scores higher in 3 categories and Qwen1.5-32B in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen1.5-32B leads 30.5 to 21.2.
- The biggest single-benchmark swing is BigCodeBench Instruct: 40.6% for Mixtral 8x22B and 32.3% for Qwen1.5-32B.
Side by side
| Mixtral 8x22B | Qwen1.5-32B | |
|---|---|---|
| Provider | Mistral AI | Alibaba (Qwen) |
| Noometry Index | 26.7 | 30.0 |
| Released | 2024-04-17 | 2024-02-04 |
| Weights | Open | Open |
| Context window | 64K | — |
| Max output | 64K | — |
| Input $ / M tokens | $2 | — |
| Output $ / M tokens | $6 | — |
| Results tracked | 34 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen1.5-32B leads
Mixtral 8x22B: 23.8 (#324), Qwen1.5-32B: 31.1 (#278)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| BigCodeBench Instruct | 40.6% | 32.3% |
| LMArena Coding | 1166 | 1155 |
| BigCodeBench Complete | 50.2% | 42% |
| WeirdML | 3.2% | — |
| HumanEval+ | 72% | — |
| MBPP+ | 64.3% | — |
Agentic & Tool Use Not comparable
Mixtral 8x22B: 22.8 (#119), Qwen1.5-32B: —
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| Cybench | 7.5% | — |
Reasoning Qwen1.5-32B leads
Mixtral 8x22B: 19.1 (#246), Qwen1.5-32B: 20.9 (#210)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| LMArena Hard Prompts | 1150 | 1131 |
| DTBench | 55.1% | — |
| Epoch Capabilities Index | 122 | — |
| ForecastBench | 56.3 | — |
Math Qwen1.5-32B leads
Mixtral 8x22B: 21.2 (#271), Qwen1.5-32B: 30.5 (#207)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| LMArena Math | 1184 | 1155 |
| Omni-MATH | 16.3% | — |
| MATH Level 5 | 24.2% | — |
Knowledge Mixtral 8x22B leads
Mixtral 8x22B: 15.1 (#288), Qwen1.5-32B: 13.5 (#291)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| GPQA Diamond | 34.1% | 30.7% |
| LMArena Expert | 1113 | 1126 |
| MMLU | 77.8% | 74.4% |
| MMLU-Pro | 46% | — |
| GPQA (HELM) | 33.4% | — |
Multilingual Mixtral 8x22B leads
Mixtral 8x22B: 32.9 (#250), Qwen1.5-32B: 31.6 (#254)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| LMArena Non-English | 1128 | 1106 |
| LMArena Chinese | 1116 | 1176 |
| LMArena French | 1165 | 1100 |
| LMArena German | 1141 | 1058 |
| LMArena Japanese | 1037 | 1027 |
| LMArena Korean | 1056 | 1007 |
| LMArena Russian | 1158 | 1073 |
| LMArena Spanish | 1150 | 1089 |
Instruction Following Too close to call
Mixtral 8x22B: 57.9 (#261), Qwen1.5-32B: 58.0 (#260)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| LMArena Instruction Following | 1147 | 1116 |
| IFEval | 72.4% | — |
Long Context Too close to call
Mixtral 8x22B: 34.6 (#242), Qwen1.5-32B: 34.7 (#241)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| LMArena Longer Query | 1144 | 1146 |
Writing & Preference Mixtral 8x22B leads
Mixtral 8x22B: 37.1 (#257), Qwen1.5-32B: 34.4 (#266)
| Benchmark | Mixtral 8x22B | Qwen1.5-32B |
|---|---|---|
| LMArena Text | 1162 | 1137 |
| LMArena Creative Writing | 1141 | 1083 |
| LMArena Multi-Turn | 1130 | 1140 |
| WildBench | 71.1% | — |
Frequently asked questions
Is Mixtral 8x22B better than Qwen1.5-32B?
Qwen1.5-32B is the stronger model overall, scoring 30.0 to 26.7 on the Noometry Index.
Is Mixtral 8x22B or Qwen1.5-32B better for coding?
Qwen1.5-32B scores higher on coding benchmarks: 31.1 versus 23.8 in the Noometry coding category.
How many benchmarks do Mixtral 8x22B and Qwen1.5-32B share?
21 benchmarks have published results for both models. Mixtral 8x22B has 34 scored results on Noometry and Qwen1.5-32B has 21.