Model comparison
MiniMax-M3 vs Mixtral 8x22B
MiniMax-M3 is the stronger model overall, scoring 43.8 to 27.1 on the Noometry Index.
Last verified . 21 shared benchmarks.
MiniMax
Mistral AISummary
- They share 21 benchmarks with published results for both. MiniMax-M3 scores higher in 8 categories and Mixtral 8x22B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 15.1.
- The biggest single-benchmark swing is GPQA Diamond: 90.9% for MiniMax-M3 and 34.1% for Mixtral 8x22B.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $2 / $6 for Mixtral 8x22B.
- MiniMax-M3 accepts more context: 1M tokens versus 64K.
Side by side
| MiniMax-M3 | Mixtral 8x22B | |
|---|---|---|
| Provider | MiniMax | Mistral AI |
| Noometry Index | 43.8 | 27.1 |
| Released | 2026-06-01 | 2024-04-17 |
| Weights | Open | Open |
| Context window | 1M | 64K |
| Max output | 512K | 64K |
| Input $ / M tokens | $0.30 | $2 |
| Output $ / M tokens | $1.20 | $6 |
| Results tracked | 41 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M3 leads
MiniMax-M3: 41.8 (#118), Mixtral 8x22B: 24.2 (#329)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Coding | 1469 | 1166 |
| FrontierCode | 14.7% | — |
| LMArena WebDev | 1482 | — |
| SciCode | 47.1% | — |
| WeirdML | — | 3.2% |
| BigCodeBench Instruct | — | 40.6% |
| BigCodeBench Complete | — | 50.2% |
| ALE-Bench | 640.02 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 64.3% |
Agentic & Tool Use Too close to call
MiniMax-M3: 22.6 (#130), Mixtral 8x22B: 23.1 (#127)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| APEX-Agents | 37.7% | — |
| OSWorld 2.0 | 4.6% | — |
| Cybench | — | 7.5% |
| GBAEval | 0.9% | — |
| Vending-Bench 2 | 2,158 | — |
Reasoning MiniMax-M3 leads
MiniMax-M3: 30.1 (#87), Mixtral 8x22B: 19.9 (#248)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1447 | 1150 |
| DTBench | 78.9% | 55.1% |
| Epoch Capabilities Index | 146.95 | 122.03 |
| ForecastBench | 61.4 | 56.3 |
| SimpleBench | 45.8% | — |
| NYT Connections (extended) | 65.1% | — |
| CritPt | 3.7% | — |
| Chess Puzzles | 14% | — |
| Mystery Game Puzzles | 8% | — |
| LMCA | 33.7% | — |
| Surface Evolver Bench | 55% | — |
Math MiniMax-M3 leads
MiniMax-M3: 40.0 (#95), Mixtral 8x22B: 22.9 (#275)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Math | 1429 | 1184 |
| OTIS Mock AIME 2024-2025 | 71.1% | — |
| ProofBench | 18% | — |
| Omni-MATH | — | 16.3% |
| MATH Level 5 | — | 24.2% |
Knowledge MiniMax-M3 leads
MiniMax-M3: 58.4 (#35), Mixtral 8x22B: 15.1 (#293)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 90.9% | 34.1% |
| LMArena Expert | 1461 | 1113 |
| MMLU-Pro | — | 46% |
| GPQA (HELM) | — | 33.4% |
| MMLU | — | 77.8% |
Multimodal Not comparable
MiniMax-M3: 40.2 (#51), Mixtral 8x22B: —
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Vision | 1253 | — |
| LMArena Document | 1435 | — |
Multilingual MiniMax-M3 leads
MiniMax-M3: 53.0 (#75), Mixtral 8x22B: 32.8 (#255)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1420 | 1128 |
| LMArena Chinese | 1463 | 1116 |
| LMArena French | 1447 | 1166 |
| LMArena German | 1426 | 1141 |
| LMArena Japanese | 1381 | 1037 |
| LMArena Korean | 1372 | 1057 |
| LMArena Russian | 1428 | 1158 |
| LMArena Spanish | 1432 | 1151 |
Instruction Following MiniMax-M3 leads
MiniMax-M3: 75.5 (#62), Mixtral 8x22B: 57.7 (#266)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Instruction Following | 1433 | 1147 |
| IFEval | — | 72.4% |
Long Context MiniMax-M3 leads
MiniMax-M3: 44.2 (#72), Mixtral 8x22B: 34.7 (#247)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1445 | 1144 |
Writing & Preference MiniMax-M3 leads
MiniMax-M3: 62.1 (#83), Mixtral 8x22B: 36.9 (#262)
| Benchmark | MiniMax-M3 | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1433 | 1162 |
| LMArena Creative Writing | 1404 | 1141 |
| LMArena Multi-Turn | 1442 | 1130 |
| WildBench | — | 71.1% |
| EQ-Bench 4 | 1150 | — |
Frequently asked questions
Is MiniMax-M3 better than Mixtral 8x22B?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 27.1 on the Noometry Index.
Which is cheaper, MiniMax-M3 or Mixtral 8x22B?
MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Mixtral 8x22B lists at $2 and $6.
Is MiniMax-M3 or Mixtral 8x22B better for coding?
MiniMax-M3 scores higher on coding benchmarks: 41.8 versus 24.2 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 64K.
How many benchmarks do MiniMax-M3 and Mixtral 8x22B share?
21 benchmarks have published results for both models. MiniMax-M3 has 41 scored results on Noometry and Mixtral 8x22B has 34.