Model comparison
Mixtral 8x22B vs Mixtral 8x7B
- Input
- $2.00 per 1M tokens
- Output
- $6.00 per 1M tokens
- Context
- 64K
- Released
- 2024-04-10
- Input
- $0.70 / 1M tokens
- Output
- $0.70 / 1M tokens
- Context
- 32K tokens
- Released
- 2023-12-09
Mixtral 8x22B scores higher on 3 of the 3 benchmarks both models report; Mixtral 8x7B is cheaper on input tokens ($0.7 vs $2 per 1M, 2.9×); Mixtral 8x22B has the larger context window (64K vs 32K).
Specs and pricing side by side
| Spec | Mixtral 8x22B | Mixtral 8x7B |
|---|---|---|
| Maker | Mistral AI | Mistral AIsame |
| Released | 2024-04-10 | 2023-12-09 |
| Context window | 64K | 32K tokens |
| Max output | 4K (la Plateforme; ~8K on some providers) | Not officially specified; bounded by the… |
| Input price | $2.00 per 1M tokens | $0.70 / 1M tokens |
| Output price | $6.00 per 1M tokens | $0.70 / 1M tokens |
| Parameters | 141B total, 39B active per token | 46.7B total parameters, 12.9B active per token (8 experts, 2 routed per token) |
| License | Apache 2.0 | Apache 2.0same |
| Weights | Open weights | Open weightssame |
| Modalities | Text | Textsame |
| Knowledge cutoff | Not publicly disclosed by Mistral AI | Not officially published by Mistral AI |
Benchmarks both models report
| Benchmark | Mixtral 8x22B | Mixtral 8x7B |
|---|---|---|
| MMLU | 77.7% | 70.6% |
| GSM8K | 90.8% | 74.4% |
| Math | 44.6% | 28.4% |
Scores are as published by each maker or its cited source; harness, reasoning effort and date can differ between rows, so treat small gaps as ties. Sources are listed on each model page.
Which one should you pick?
Pick Mixtral 8x22B if…
- You care about MMLU, GSM8K, Math (it scores higher on 3 shared benchmarks).
- You need the longer context (64K vs 32K).
- Your use case is: self-hosted open-weight LLM deployments where data must stay on-premises or in a controlled environment.
Pick Mixtral 8x7B if…
- Cost matters: input tokens cost 2.9× less ($0.7 vs $2 per 1M).
- Your use case is: self-hosted chat assistants and RAG backends where Apache 2.0 licensing matters.
FAQ
Which is better, Mixtral 8x22B or Mixtral 8x7B?
Mixtral 8x22B scores higher on 3 of the 3 benchmarks both models report; Mixtral 8x7B is cheaper on input tokens ($0.7 vs $2 per 1M, 2.9×); Mixtral 8x22B has the larger context window (64K vs 32K). The better pick depends on whether you value benchmark scores, price per token, context length or openness for your workload.
How much cheaper is Mixtral 8x7B?
Mixtral 8x22B: $2.00 per 1M tokens input, $6.00 per 1M tokens output. Mixtral 8x7B: $0.70 / 1M tokens input, $0.70 / 1M tokens output (list API prices as published by each maker).
What are the context windows of Mixtral 8x22B and Mixtral 8x7B?
Mixtral 8x22B offers 64K with up to 4K (la Plateforme; ~8K on some providers) output; Mixtral 8x7B offers 32K tokens with up to Not officially specified; bounded by the 32K-token context window output.
Is Mixtral 8x22B newer than Mixtral 8x7B?
Mixtral 8x22B was released 2024-04-10; Mixtral 8x7B was released 2023-12-09.