AI tools for MoE

Edge0 is a streaming inference framework for running mixture-of-experts language models with a small active-memory footprint. It streams expert weights from SSD storage, allowin

Ling-3.0-flash-base is an open base language model from inclusionAI for continued pretraining, mid-training, fine-tuning, and research. It is a 124B-total-parameter sparse mixtu

Qwen3.8-2.4T-A95B is a very large mixture-of-experts language model with 2.4 trillion total parameters and 95 billion activated parameters. It targets advanced coding, agentic w

Instella-MoE is AMD's fully open Mixture-of-Experts language model with 16 billion total parameters and about 2.8 billion active parameters per token. It was trained from scratc

ZAYA1-8B is Zyphra's compact mixture-of-experts language model built for high intelligence density at a small active-parameter budget. The model is designed to perform strongly

Uni-MoE-2.0-Omni is an advanced omnimodal large model built on the powerful Qwen2.5-7B core. It represents a significant rebuild and improvement over previous versions, incorpor

Qwen3-VL is the latest multimodal large language model series developed by the Qwen team at Alibaba Cloud. It represents the most powerful vision-language model in the Qwen seri

Qwen3-Omni is a cutting-edge multilingual omni-modal foundation model developed to natively process a wide range of inputs including text, images, audio, and video. It delivers

Qwen3-Max is Alibaba's largest and most advanced large language model released in January 2025. It features over 1 trillion parameters and has been trained on an enormous datase

Wan2.2 is an open and advanced large-scale video generative model that introduces a Mixture-of-Experts (MoE) architecture into video diffusion models. This architecture separate

Bytedance Seed 1.5-VL is a powerful and efficient vision-language foundation model designed for advanced general-purpose multimodal understanding and reasoning. It achieves top

Qwen3 is the latest generation of large language models developed by Alibaba Group, designed to deliver exceptional performance across a wide range of reasoning, language unders

ICEdit is an innovative, open-source framework for instruction-based image editing that dramatically simplifies the process of modifying images using natural language commands.

Qwen2.5-Max is Alibaba's most advanced large language model, engineered to compete with leading models such as GPT-4o, Claude 3.5 Sonnet, and DeepSeek V3. Developed using a Mixt

DeepSeek-R1 is a state-of-the-art large language model developed by DeepSeek AI, designed to set new benchmarks in reasoning, mathematics, coding, and complex problem-solving. L

ChatLaw is an advanced AI legal assistant designed to enhance the efficiency and reliability of legal services through a multi-agent collaborative framework. Utilizing a Mixture