NaiveAI releases open-weight coding model with one-million-token context

NaiveAI released the weights and inference code for Naive-N0.5-Flash under the MIT license on September 27, TechBooky reported. The model is designed for coding and AI research and supports a one-million-token context window, allowing it to process large inputs. Developers can inspect and adapt the model instead of relying solely on a hosted service. NaiveAI says AI systems helped write code and run experiments during development, while human researchers directed the work and made key decisions. The model card says the weights alone require about 315GB of storage and recommends Nvidia GPUs capable of FP8 computation, limiting practical self-hosting to well-equipped users.