M3 Chat

M3 Chat is an OSS AI chat service that's completely free, private, actively maintained and built with Next.js

Why use M3 Chat?

Most AI chat web app services are slow, expensive, or require accounts (with heavy message limits). M3 Chat is an open-source, free, no account required, AI chat web app with unlimited messages. I'm currently running all models on a Mac Studio M3 Ultra (which is why it's named M3 chat) using Ollama.

Models

• LLaMA 3 8B (General Purpose) — llama3:8b
• LLaMA 2 (Uncensored) — llama2-uncensored
• Gemma 3 (General Purpose) — gemma3
• Gemma (General Purpose) — gemma
• Phi-3 Mini (Fastest) — phi3:mini
• Mistral 7B (Balanced) — mistral
• OpenCoder 8B (Code) - opencoder:8b
• Gemma 2B (Tiny + Reasonable) — gemma:2b
• Gemma 7B (Chat Capable) — gemma:7b
• Qwen 1.5 7B (Chat & Reasoning) — qwen:7b
• Qwen 2.5 Coder (Code) — qwen2.5-coder
• Qwen 3 (General Purpose) — qwen3
• DeepSeek Coder 6.7B (Code) — deepseek-coder:6.7b
• DeepSeek V2 16B (General Purpose) — deepseek-v2:16b
• Dolphin Mistral 7B (Uncensored, Code) — dolphin-mistral:7b
• Dolphin 3 8B (General Purpose) — dolphin3
• StarCoder2 7B (Coding) — starcoder2:7b
• Magistral (Reasoning) — magistral
• Devstral (Code) — devstral

Contributing

M3 Chat is open to contributions, suggestions, or any feedback. Please follow our CONTRIBUTING.md guide.