Meissa: Multi-modal Medical Agentic Intelligence
What it is
Meissa is a 4B-parameter multi-modal medical LLM designed to operate fully offline with agentic capabilities. It uses trajectory-based distillation to learn when and how to engage in tool use and multi-agent collaboration, matching or exceeding proprietary frontier agents on 10 of 16 medical benchmarks.
The problem it solves
High cost, latency, and privacy risks associated with deploying proprietary frontier models (GPT, Gemini) for medical agentic tasks.
What you could use it for
- Offline medical image understanding and clinical reasoning
- Radiology VQA with continuous tool calling
- Pathology analysis with interleaved thinking and image injection
- Multi-agent debate for complex clinical questions
- Multi-turn doctor-patient simulation with exam ordering
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.