Tag
Every Latent Daily story, tool, model and resource tagged “Coding”.
Meta published Muse Spark Open, a mixture-of-experts model with roughly 816B parameters, 42B active per token, multimodal input and a 1M-token context window.
AI Desk1 min read
Reflection, a new US lab, put out Beam: a 501B mixture-of-experts model with 23B active parameters, a 1M-token context and an Apache 2.0 license.
AI Desk1 min read
In a controlled experiment, developers with GitHub Copilot completed a coding task about 56% faster. The study design, results, limitations and lessons for teams.
AI Desk3 min read
An open-weight 501B mixture-of-experts model from Reflection for coding and agentic tasks, with a 1M-token context window.
Meta's open-weight multimodal mixture-of-experts model, trained for agentic coding, with roughly 816B parameters and a 1M-token context.
Alibaba's open-weight 32B coding model, one of the strongest open code models of its generation, released under Apache 2.0.
Mistral AI's first code model: 22 billion parameters trained on 80+ programming languages, with fill-in-the-middle support for completion.