Abstract
Micius-Flash is a 27B-parameter multimodal reasoning model trained with a knowledge-tree data pipeline and multi-teacher on-policy distillation, achieving strong hardware and semiconductor agent benchmark results and supporting tool-integrated engineering workflows.
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 synthesis, and quality validation with hard-example augmentation. From this data, we train expert models for semiconductor knowledge and reasoning, hardware coding, and semiconductor multimodal understanding, along with agentic experts. Subsequently, we employ Multi-Teacher On-Policy Distillation (MOPD) to supervise training on sequences generated by Micius itself, thereby integrating the capabilities of these expert models into a unified model. On more than 20 datasets with about 15K records, Micius-Flash averages 64.2% on hardware code benchmarks, 15.0 points above a same-size baseline. It scores 62.4% on a semiconductor-specific agent benchmark, exceeding the same-size baseline by 17.2 points. We further develop Micius Harness for industrial applications, connecting the model to specialized tools, including EDA tools and KiCad, as well as execution environments. Case studies in PCB design, chip design workflow automation, and software development demonstrate the model's ability to use tool execution feedback for multi-step planning, code debugging, and iterative verification. We release Micius as an open-source model to advance the development of agents for science and engineering.
Community
We present Micius-Flash, an open-source 27B multimodal model designed for semiconductor engineering and scientific applications. Micius combines a knowledge-tree-driven data pipeline, specialized expert models, and Multi-Teacher On-Policy Distillation (MOPD) to unify semiconductor knowledge, hardware coding, multimodal understanding, and agentic capabilities. It achieves 64.2% average accuracy on hardware coding benchmarks and 62.4% on a semiconductor-specific agent benchmark, outperforming same-size baselines by 15.0 and 17.2 percentage points, respectively. We also introduce Micius Harness, enabling tool-augmented workflows with EDA tools, KiCad, and execution environments. Our case studies demonstrate practical capabilities in PCB design, chip design automation, and software development. We release Micius-Flash and aim to advance open-source AI agents for science and engineering.
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper