mistral-7b-instruct vs open-mixtral-8x7b

The mistralai/mistral-7b-instruct series is a 7B-parameter language model fine-tuned for instruction-based tasks. It supports an extended context window (up to 32K tokens) and can handle function calling, demonstrating strong instruct performance. As an early demonstration, it lacks built-in content moderation mechanisms.

Mixtral 8x7B is a high-quality sparse mixture of experts (SMoE) large language model with open weights, released under Apache 2.0 license. Despite having 45 billion parameters, its architecture requires compute equivalent to a 14 billion parameter model, enabling 6x faster inference and strong performance that outperforms Llama 2 70B and matches or exceeds GPT-3.5 on many benchmarks.