Model comparison
Mixtral 8x22B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 45.3 to 26.7 on the Noometry Index.
Last verified . 19 shared benchmarks.
Mistral AI
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Mixtral 8x22B scores higher in 0 categories and Qwen3.8 27B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.8 27B leads 65.9 to 37.1.
- The biggest single-benchmark swing is DTBench: 55.1% for Mixtral 8x22B and 88% for Qwen3.8 27B.
- Qwen3.8 27B is cheaper at $0.99 / $1.49 per million input/output tokens, against $2 / $6 for Mixtral 8x22B.
- Qwen3.8 27B accepts more context: 262K tokens versus 64K.
Side by side
| Mixtral 8x22B | Qwen3.8 27B | |
|---|---|---|
| Provider | Mistral AI | Alibaba (Qwen) |
| Noometry Index | 26.7 | 45.3 |
| Released | 2024-04-17 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 64K | 262K |
| Max output | 64K | 33K |
| Input $ / M tokens | $2 | $0.99 |
| Output $ / M tokens | $6 | $1.49 |
| Results tracked | 34 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Mixtral 8x22B: 23.8 (#324), Qwen3.8 27B: 50.3 (#42)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1166 | 1487 |
| LMArena WebDev | — | 1592 |
| SciCode | — | 46.6% |
| WeirdML | 3.2% | — |
| BigCodeBench Instruct | 40.6% | — |
| BigCodeBench Complete | 50.2% | — |
| HumanEval+ | 72% | — |
| MBPP+ | 64.3% | — |
Agentic & Tool Use Not comparable
Mixtral 8x22B: 22.8 (#119), Qwen3.8 27B: —
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| Cybench | 7.5% | — |
Reasoning Qwen3.8 27B leads
Mixtral 8x22B: 19.1 (#246), Qwen3.8 27B: 31.5 (#71)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1150 | 1459 |
| DTBench | 55.1% | 88% |
| Epoch Capabilities Index | 122 | 149.38 |
| NYT Connections (extended) | — | 54.5% |
| CritPt | — | 5.4% |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| ForecastBench | 56.3 | — |
Math Qwen3.8 27B leads
Mixtral 8x22B: 21.2 (#271), Qwen3.8 27B: 35.7 (#139)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1184 | 1453 |
| ProofBench | — | 16% |
| Omni-MATH | 16.3% | — |
| MATH Level 5 | 24.2% | — |
Knowledge Qwen3.8 27B leads
Mixtral 8x22B: 15.1 (#288), Qwen3.8 27B: 40.9 (#105)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1113 | 1482 |
| GPQA Diamond | 34.1% | — |
| MMLU-Pro | 46% | — |
| GPQA (HELM) | 33.4% | — |
| MMLU | 77.8% | — |
Multimodal Not comparable
Mixtral 8x22B: —, Qwen3.8 27B: 40.5 (#36)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1270 |
Multilingual Qwen3.8 27B leads
Mixtral 8x22B: 32.9 (#250), Qwen3.8 27B: 54.0 (#54)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1128 | 1431 |
| LMArena Chinese | 1116 | 1504 |
| LMArena French | 1165 | 1468 |
| LMArena German | 1141 | 1435 |
| LMArena Japanese | 1037 | 1392 |
| LMArena Korean | 1056 | 1389 |
| LMArena Russian | 1158 | 1421 |
| LMArena Spanish | 1150 | 1444 |
Instruction Following Qwen3.8 27B leads
Mixtral 8x22B: 57.9 (#261), Qwen3.8 27B: 75.8 (#54)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1147 | 1436 |
| IFEval | 72.4% | — |
Long Context Qwen3.8 27B leads
Mixtral 8x22B: 34.6 (#242), Qwen3.8 27B: 44.3 (#66)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1144 | 1448 |
Writing & Preference Qwen3.8 27B leads
Mixtral 8x22B: 37.1 (#257), Qwen3.8 27B: 65.9 (#41)
| Benchmark | Mixtral 8x22B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1162 | 1441 |
| LMArena Creative Writing | 1141 | 1382 |
| LMArena Multi-Turn | 1130 | 1443 |
| EQ-Bench Creative Writing | — | 1671 |
| WildBench | 71.1% | — |
Frequently asked questions
Is Mixtral 8x22B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 45.3 to 26.7 on the Noometry Index.
Which is cheaper, Mixtral 8x22B or Qwen3.8 27B?
Qwen3.8 27B is cheaper. It lists at $0.99 per million input tokens and $1.49 per million output tokens; Mixtral 8x22B lists at $2 and $6.
Is Mixtral 8x22B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.3 versus 23.8 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 64K.
How many benchmarks do Mixtral 8x22B and Qwen3.8 27B share?
19 benchmarks have published results for both models. Mixtral 8x22B has 34 scored results on Noometry and Qwen3.8 27B has 28.