Tremendous excited GPT-5.5 is rolling out to GitHub Copilot, M365 Copilot, Copilot Studio, and Foundry right this moment.
With deeper reasoning, stronger multistep execution, and higher efficiency throughout lengthy, complicated duties, GPT-5.5 helps you go from concept to execution sooner with fewer iterations to get to the correct consequence.
It’s all about serving to you select the correct mannequin, or fashions, for the correct activity throughout your workflow.
In GitHub Copilot CLI, you’ll be able to transfer throughout fashions relying on the job.
Use sooner fashions to scaffold and discover rapidly. Use deeper reasoning fashions to plan, perceive necessities, and make the correct tradeoffs.
Then GPT-5.5 can take that plan and switch it into working code, iterating, resolving errors, utilizing instruments, and validating outcomes alongside the way in which.
And with our Rubber Duck agent you get a multi-model reflection loop as a reviewer. GPT-5.5 can assessment the output of one other mannequin or vice versa, as seen in Seth Juarez‘s video under!
These similar capabilities additionally present up in on a regular basis work.
In M365 Copilot, we help a number of fashions together with GPT-5.5 with Work IQ to purpose over your work information, execute complicated duties, and generate full paperwork and spreadsheets.
And in Foundry, you’ll be able to construct with GPT-5.5 utilizing the safety, governance, and orchestration required for enterprise functions and workflows.
Wanting ahead to what you construct!
Transcript
Transcript
Transcript
GPD 5.5 has arrived in GitHub Copilot and let me inform you, this model new mannequin is constructed for severe coding work. Let me present you what that appears like by including an actual function to the app I exploit to construct demo storyboards. My boss simply despatched a function request by way of e-mail, so as an alternative of copy pasting context, I exploit Work IQ to tug that message immediately into the GitHub Copilot CLI and switch it into actionable product necessities from their copilot. Creates a function department, explores the code base, implements the readiness logic. UI and provides targeted exams for the brand new conduct. Then GD 5.5 rubber geese its personal implementation by bringing in two reviewers, Claude Sonnet 4.6 and GT 5.4. That assessment catches 2 significant gaps and GD 5.5 fixes them earlier than transport. By the top you see TBD 5.5 run the total developer loop, purpose over actual work, IQ context proper and validate code, incorporate multi mannequin critique, open a PR and launch the app for finish to finish testing. To wrap issues up, here is a brand new function in motion. It turned out to be an ideal addition and I did not even should open the e-mail myself. Win win.