Over the past two months, we’ve been previewing a new capability within Copilot Studio that lets you build agents capable of taking on more complex business processes. Today, we’re excited to share that this capability is now generally available for production use and to introduce its new name: the GitHub Copilot harness in Copilot Studio . The GitHub Copilot harness gives Copilot Studio the coding and reasoning capabilities behind our most advanced agent experiences (like Copilot Cowork and the GitHub Copilot coding agent ). This means your agents can now handle work that used to be out of reach, including processes that have many steps, many sources, and ambiguous decision points. The new harness is built for complex, long-horizon work using the latest frontier reasoning models like Opus 5, GPT-5.6 Sol, and Fable 5. It can plan, reason through dynamic problems, run an agentic loop, use skills, integrate workflows, connect to tools and agents in other platforms, and produce rich, multi-part outputs. In our own testing with real world business process evals, Copilot Studio shows significant performance and quality gains when using the GitHub Copilot harness. Improvements in multi-tool use, file analysis, code analysis, and knowledge quality mean these agents are more capable of taking on your most complex business processes. To help you build these more capable agents, we’re also delivering new experiences for makers building agents with the GitHub Copilot harness. The agent designer is more intuitive for authoring, putting the most important tools right within reach to make authoring faster, while retaining full agent lifecycle management features. Then, the w orkflow designer gives a visual canvas to understand and edit workflows, including adding agent nodes and running workflow evals. And soon, natural language authoring will let you describe your business goal and assemble the right combination of agents and workflows through a multi-turn conversation. Agents running on the GitHub Copilot harness use usage-based billing for all work, regardless of Microsoft 365 Copilot licensing. You pay for your agent’s usage based on the models you choose, the organizational context and tools you add, and runtime used. Certain AI-driven maker experiences, like natural language authoring, evaluations, and testing, will also fall under usage-based billing if building with the GitHub Copilot harness. You can learn more about Copilot Stu dio usag e-based billing her e .

Adding a new harness does not change the ones you already rely on. Different harnesses can be optimized for different outcomes, so rather than relying on a single one-size-fits-all approach, Copilot Studio now lets you choose the harness that best fits your scenario. This means Copilot Studio supports three harnesses today: The Copilot Chat harness, which uses the same harness as Microsoft 365 Copilot Chat and is ideal for customizing Copilot Chat experiences. The Standard harness, which most agents built in Copilot Studio use today and is great for conversational agents with rules-based topics. The new GitHub Copilot harness, which uses the power of GitHub Copilot SDK to automate complex, agentic business processes. We will continue to support the Copilot Chat and Standard harnesses in Copilot Studio for both existing agents and authoring of new agents. Microsoft 365 Copilot licensed users continue to benefit from fair use of Copilot Chat or Standard harness agents included in their Microsoft 365 Copilot license. For all other usage, the Copilot Chat and Standard harnesses will continue to be billed using the existing fixed rate card. The GitHub Copilot harness is generally available now, and you can start creating these agents directly from the Copilot Studio homepage. We can’t wait to see what you build. Share your thoughts using the feedback control in the top right corner of the product. Your input directly shapes what we build—and how you build—next. Try Copilot Studio today!