Llama 3 vs LLaMA

Phi-3, a family of open AI models developed by Microsoft. Phi-3 models are the most capable and cost-effective small language models (SLMs) available, outperforming models of the same size and next size up across a variety of language, reasoning, coding, and math benchmarks.

DeepSeek's 4th generation flagship model. V4-Pro (1.6T MoE) and V4-Flash (284B MoE) featuring 1M context, native multimodality, and Hybrid Attention for extreme efficiency.

Mixtral 8x7B, a high-quality sparse mixture of experts model (SMoE) with open weights. Mixtral outperforms Llama 2 70B on most benchmarks with 6x faster inference. It matches or outperforms GPT3.5 on most standard benchmarks. paper：https://arxiv.org/pdf/2401.04088.pdf news：https://mistral.ai/news/mixtral-of-experts/

A strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token.