MetaOct 15, 2026
Muse Spark Open
Meta's open-weight multimodal mixture-of-experts model, trained for agentic coding, with roughly 816B parameters and a 1M-token context.
- Type
- Open-Weight
- Context
- 1M
- Weights
- Open
Tools, Software & Models / Models Directory
Models whose weights you can download and run.
Models directory
Meta's open-weight multimodal mixture-of-experts model, trained for agentic coding, with roughly 816B parameters and a 1M-token context.
An open-weight 501B mixture-of-experts model from Reflection for coding and agentic tasks, with a 1M-token context window.
When should you self-host an open model like Llama, Mistral or Qwen, and when is a closed API better? Cost, quality, privacy, control and licensing compared.
AI Desk4 min read
When fine-tuning helps, how LoRA and QLoRA make it affordable, how to prepare data, train, evaluate and deploy an adapter — and when not to fine-tune at all.
AI Desk3 min read