Faster, with the same price per token
A change, a reply, another comment, another wait. In that kind of workflow, model speed starts to matter. Anthropic introduced Sonnet 5.5 on September 28, 2026, claiming response generation more than 30 percent faster than Sonnet 5. Typical tasks should cost up to 30 percent less through lower token usage.
Standard API rates remain $2 per million input tokens and $10 per million output tokens. The savings therefore concern claimed consumption during a job. Anthropic’s listed uses include fixing bugs and preparing documents, presentations and spreadsheets.
Small changes can still take the whole morning
What interests us most is the rhythm of everyday work. You refine a form, shorten a change description or organize meeting notes. The result needs a few more edits, and you want to keep the topic in your head. Shorter waits between replies may help.
With meeting notes, start with something concrete: decisions, tasks with owners and unanswered questions. Add a simple rule: “If no deadline was agreed in the conversation, leave the field empty.” You can immediately see what to hand to the team and what still needs a follow-up.
Show an example of what you want back
“Make it professional” sounds familiar, but what should the result actually look like? Attach a note you wrote, an issue description or a finished table instead. Specify the order of information, length and number format. A short example says more about your expectations.
You will probably edit the spreadsheet or presentation again. Check that you can change a number, heading and chart without rebuilding everything. The first review rarely settles it; you should be able to incorporate the team’s comments in the delivered file.
Sources and further reading
The sources behind this article. Instructions and example scenarios are suggestions from our editorial team.