1/ In GitHub Copilot CLI, you can move across models depending on the job. Use faster models to scaffold and explore quickly. Use deeper reasoning models to plan, understand requirements, and make the right tradeoffs. Then GPT-5.5 can take that plan and turn it into working code, iterating, resolving... errors, using tools, and validating results along the way.show more

Super excited GPT-5.5 is rolling out to GitHub Copilot, M365 Copilot, Copilot Studio, and Foundry today. With deeper reasoning, stronger multistep execution, and better performance across long, complex tasks, GPT-5.5 helps you go from idea to execution faster with fewer iterations to get to the right outcome. It’s all about helping you choose the right model, or models, for the right task across your workflow.

2/ And with our Rubber Duck agent you get a multi-model reflection loop as a reviewer. GPT-5.5 can review the output of another model or vice versa!

3/ These same capabilities also show up in everyday work. In M365 Copilot, we support multiple models including GPT-5.5 with Work IQ to reason over your work data, execute complex tasks, and generate complete documents and spreadsheets. And in Foundry, you can build with GPT-5.5 using the security, governance, and orchestration required for enterprise applications and workflows. Looking forward to what you build!

This is where agent architecture is heading too. you dont want one model doing everything, you want fast cheap models for routine checks and heavy reasoning models for actual decisions. the cost difference between running haiku for monitoring vs opus for strategy is like 50x. multi model isnt a feature, its the only way agents become affordable at scale

You use macOS😅 Even Microsoft employee doesn’t use garbage windows.

Github Copilot CLI was how to keep your finger on the pulse of frontier models like Opus, then you took it away and after 3 years that money was better spent upgrading ChatGPT to $100/mth 😢

Any time any model any where? Really?

this is the autonomy ladder. who assigns which rung to each workflow is the PM job, not the model's

Really interesting point. I’m building an AI that helps people simulate life decisions (career, moving, etc.) What I’ve seen: People don’t need more advice — they need clarity on outcomes. Still early, but promising:

How to test it, if the selecting plans is disabled????

@gitlawb can do that much better but guess you have not heard or explored it Your company should look at it and get in touch with @gitlawb founder @kevincodex