Microsoft Research has introduced Orchard, an open-source framework designed to reduce the costs and complexity associated with training autonomous AI agents. This framework aims to democratize access to state-of-the-art agentic AI systems by providing reusable infrastructure components, eliminating the need for researchers to build custom environments from scratch. Orchard supports various tasks such as coding, web browsing, and tool usage, and includes pre-built training recipes for software engineering, graphical user interface navigation, and productivity tasks. AI
IMPACT This framework aims to lower the barrier to entry for developing and training AI agents, potentially accelerating research and application development in agentic AI.
RANK_REASON This is a framework release from a research division of a large tech company, not a core frontier model release.
- Google Cloud
- Kubernetes
- Microsoft
- Microsoft Research
- Orchard
- Orchard-Claw
- Orchard Env
- Orchard-GUI
- Orchard-SWE
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
