Back to feed
Mistral AIModel ReleaseDec 112 min read

Mistral AI released Mixtral 8x7B, a high-quality sparse Mixture of Experts open weights model

HIGHLIGHTS
  • Mistral AI released Mixtral 8x7B, a high-quality sparse Mixture of Experts open weights model
  • It outperforms Llama 2 70B on most benchmarks with 6x faster inference, routing each token to 2 active expert networks.
  • This is significant as it offers improved performance and efficiency in AI model benchmarks.
Published:
Official Source (Official Blog)

Related Activities