A 27-billion-parameter engineering model called Micius-Flash was introduced with native multimodal input, deep reasoning and a 256K-token context window. The model is built on a Micius data pipeline described as a knowledge-tree-driven data flywheel with three stages, beginning with knowledge tree construction and expansion. The release targets engineering workloads with a lightweight agentic design.
We introduce Micius-Flash, a 27-billion-parameter model for engineering with native multimodal input, deep reasoning and a 256K-token context window. We design the Micius data pipeline as a knowledge-tree-driven data flywheel comprising three stages: knowledge tree construction and expansion, data s
EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.