Mistral AI
Released: 2024-04-10
Mixtral 8x22B
Model Specifications
Context Window 64k tokens
Parameters 141B
Pricing (Input) $2.00 / M tokens
Pricing (Output) $6.00 / M tokens
What is Mixtral 8x22B?
Mixtral 8x22B is Mistral AI’s largest open-weight Mixture of Experts (MoE) model. Released in April 2024, it features 141 billion parameters (activating 39 billion parameters per token).
Licensed under Apache 2.0, it offers a strong combination of open flexibility and high intelligence for enterprise tasks.
Key Capabilities
- MoE parameter efficiency: Fast generation speeds despite its parameter count.
- Apache 2.0 License: Fully open for commercial deployment and fine-tuning.
- Multilingual depth: Strong performance across multiple European languages.
Ideal Use Cases
- Custom model fine-tuning: Serving as a base for custom domain models.
- Enterprise data extraction: Parsing data from unstructured sources.
- Translation services: Handling content across languages.
Historical figures, architectures, and capabilities are for informational purposes only. Not technical, professional, legal, or financial advice. Sources: Benchmark evaluations derived from public developer statements.