Meta's open-weight 816B-parameter mixture-of-experts model (about 42B active), trained for agentic coding, with a 1M-token context window. Supports image input, tool calling, structured output, and reasoning effort.
Meta's open-weight 816B-parameter mixture-of-experts model (about 42B active), trained for agentic coding, with a 1M-token context window. Supports image input, tool calling, structured output, and reasoning effort.