Claude Opus 5.5 Is Now Live: The Newest Opus for Your Team

Update, October 2026: thinking power is now Light, Standard or Expert, and we pick the model behind each, so the five choices below are no longer in the picker. Light uses half the credits of Standard, Expert twice as many, and Free runs on Standard. Running your employees on your own Claude or ChatGPT plan is part of the Enterprise plan.

Claude Opus 5.5, the newest Opus, is now available to every AgentTeams employee that runs on Claude. It is a step up from Opus 5. Since November 2026 it is the Standard thinking power, the default for every employee on Claude, so nothing needs choosing.

If you read our post on matching thinking power to the work, you know how we think about a new mind like this one. It is not an upgrade to switch on everywhere. It is a stronger option for the employees whose work asks for more.

Where Opus 5.5 fits

Your team now has five thinking powers on the Claude side, from deepest to fastest:

• Fable 5.2 is Expert, the deepest thinker, for the hardest work: contract reviews, numbers that refuse to match, the case five people already gave up on.
• Opus 5.5 is Standard, the default: a strong fit for most roles and demanding everyday work.
• Haiku 5.5 is Light: the fastest replies at the lowest cost.

When Fable 5.1 arrived, we said the point was the right mind for each job, not the biggest mind for every job. Opus 5.5 is that default now, with Fable 5.2 one step above it, for the work that is more demanding than most.

How to try it

Every employee has a Thinking power setting on their Settings tab. Pick Opus 5.5 there and it applies from their next conversation. To move the whole organization at once, change the default in your organization's Settings, where you connected Claude.

The pattern that works: pick one employee whose work is genuinely demanding, such as long analyses, multi-step research, or writing that has to be right the first time. Move them to Opus 5.5 for a week and compare the output with the week before. Then decide with evidence, not with a benchmark.

Teams on ChatGPT

Opus 5.5 is on the Claude side. Teams that run their employees through a ChatGPT plan keep GPT-6 Astra and the rest of that lineup. Wondering which side suits you? Our comparison of running AI agents on Claude or ChatGPT walks through it.

This is one of the quiet advantages of how AgentTeams is built: your employees are not tied to one mind or one vendor. When a better one appears, it shows up in the picker and your team can start using it the same week.