Model comparison
Gemma 4 31B IT vs Mixtral 8x22B
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 27.1 on the Noometry Index.
Last verified . 18 shared benchmarks.
Google
Mistral AISummary
- They share 18 benchmarks with published results for both. Gemma 4 31B IT scores higher in 8 categories and Mixtral 8x22B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 4 31B IT leads 60.5 to 36.9.
- The biggest single-benchmark swing is WeirdML: 52.3% for Gemma 4 31B IT and 3.2% for Mixtral 8x22B.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $2 / $6 for Mixtral 8x22B.
- Gemma 4 31B IT accepts more context: 262K tokens versus 64K.
Side by side
| Gemma 4 31B IT | Mixtral 8x22B | |
|---|---|---|
| Provider | Mistral AI | |
| Noometry Index | 43.5 | 27.1 |
| Released | 2026-04-02 | 2024-04-17 |
| Weights | Open | Open |
| Context window | 262K | 64K |
| Max output | 33K | 64K |
| Input $ / M tokens | $0.09 | $2 |
| Output $ / M tokens | $0.34 | $6 |
| Results tracked | 35 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 4 31B IT leads
Gemma 4 31B IT: 42.3 (#108), Mixtral 8x22B: 24.2 (#329)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| WeirdML | 52.3% | 3.2% |
| LMArena Coding | 1459 | 1166 |
| LMArena WebDev | 1366 | — |
| SciCode | 43.4% | — |
| BigCodeBench Instruct | — | 40.6% |
| BigCodeBench Complete | — | 50.2% |
| ALE-Bench | 925.5 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 64.3% |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, Mixtral 8x22B: 23.1 (#127)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| Cybench | — | 7.5% |
Reasoning Gemma 4 31B IT leads
Gemma 4 31B IT: 27.2 (#122), Mixtral 8x22B: 19.9 (#248)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1448 | 1150 |
| DTBench | 82.7% | 55.1% |
| Epoch Capabilities Index | 142.74 | 122.03 |
| Kagi LLM Benchmark | 63.5% | — |
| NYT Connections (extended) | 70.6% | — |
| CritPt | 1.4% | — |
| Chess Puzzles | 5% | — |
| Thematic Generalization | 53% | — |
| LMCA | 39.3% | — |
| Surface Evolver Bench | 30.6% | — |
| ForecastBench | — | 56.3 |
Math Gemma 4 31B IT leads
Gemma 4 31B IT: 43.2 (#81), Mixtral 8x22B: 22.9 (#275)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Math | 1465 | 1184 |
| OTIS Mock AIME 2024-2025 | 73.3% | — |
| Omni-MATH | — | 16.3% |
| MATH Level 5 | — | 24.2% |
Knowledge Gemma 4 31B IT leads
Gemma 4 31B IT: 37.9 (#151), Mixtral 8x22B: 15.1 (#293)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 75.8% | 34.1% |
| LMArena Expert | 1465 | 1113 |
| SimpleQA Verified | 10.4% | — |
| MMLU-Pro | — | 46% |
| Vectara Hallucination Rate | 7.4% | — |
| GPQA (HELM) | — | 33.4% |
| MMLU | — | 77.8% |
Multimodal Not comparable
Gemma 4 31B IT: 41.6 (#34), Mixtral 8x22B: —
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Vision | 1277 | — |
| LMArena Document | 1425 | — |
Multilingual Gemma 4 31B IT leads
Gemma 4 31B IT: 53.8 (#57), Mixtral 8x22B: 32.8 (#255)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1431 | 1128 |
| LMArena Chinese | 1476 | 1116 |
| LMArena French | 1435 | 1166 |
| LMArena Russian | 1460 | 1158 |
| LMArena Spanish | 1444 | 1151 |
| LMArena German | — | 1141 |
| LMArena Japanese | — | 1037 |
| LMArena Korean | — | 1057 |
Instruction Following Gemma 4 31B IT leads
Gemma 4 31B IT: 75.5 (#61), Mixtral 8x22B: 57.7 (#266)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Instruction Following | 1433 | 1147 |
| IFEval | — | 72.4% |
Long Context Gemma 4 31B IT leads
Gemma 4 31B IT: 44.2 (#71), Mixtral 8x22B: 34.7 (#247)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1446 | 1144 |
Writing & Preference Gemma 4 31B IT leads
Gemma 4 31B IT: 60.5 (#96), Mixtral 8x22B: 36.9 (#262)
| Benchmark | Gemma 4 31B IT | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1443 | 1162 |
| LMArena Creative Writing | 1415 | 1141 |
| LMArena Multi-Turn | 1452 | 1130 |
| EQ-Bench Creative Writing | 1368 | — |
| WildBench | — | 71.1% |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than Mixtral 8x22B?
Gemma 4 31B IT is the stronger model overall, scoring 43.5 to 27.1 on the Noometry Index.
Which is cheaper, Gemma 4 31B IT or Mixtral 8x22B?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; Mixtral 8x22B lists at $2 and $6.
Is Gemma 4 31B IT or Mixtral 8x22B better for coding?
Gemma 4 31B IT scores higher on coding benchmarks: 42.3 versus 24.2 in the Noometry coding category.
Which has the bigger context window?
Gemma 4 31B IT does, with 262K tokens against 64K.
How many benchmarks do Gemma 4 31B IT and Mixtral 8x22B share?
18 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and Mixtral 8x22B has 34.