When you fix what Claude made for you little by little in conversation, the time spent waiting for each reply adds up. Even for small fixes like "just the button color" or "just this one sentence," long waits tend to make asking again and again feel like a chore.
"Claude Sonnet 5.5," a new model announced on September 28, 2026 by Anthropic, the company that makes Claude, is a model that shortens this wait. A model is the part that acts as Claude's brain.
Let's look at what changed, why it helps people starting from zero, and how to ask in a way that makes the most of this speed.
Contents
What changed: 30%+ faster than the previous Sonnet, with clearer strengths
According to Anthropic's announcement, Sonnet 5.5 generates replies 30%+ faster than the previous model, Sonnet 5, and the cost of doing the same work is up to 30% lower in most cases. Compared with Opus 5.5, the higher-tier model released around the same time, it's positioned as faster and easier to use casually.
The announcement also spells out what each model is good at. Opus 5.5 is for complex work that requires careful judgment. Sonnet 5.5 is strongest at well-scoped everyday tasks, fixing bugs, and creating documents, slides, and spreadsheets, and is said to have a good eye for design.
Sonnet 5.5 is a model that's good at getting clearly defined work done quickly.
Why it helps people starting from zero
When you build something from zero, you usually go back and forth, looking, fixing, and looking again, rather than finishing it in one go. The announcement explains that Sonnet 5.5's speed suits fixing less complex things quickly, over and over.
It also shares feedback from early testers that it's a better partner for working together than Sonnet 5 and a more natural conversational partner. Some also noted that it's good at adding polish to how screens look and at making decks that follow a slide template.
When each wait is shorter, you don't have to hesitate to ask for small fixes. Being able to casually ask "just fix this part" again and again is the biggest change.
Your first request: once the direction is set, stack small fixes at a good pace
To make the most of Sonnet 5.5's speed, it's a good idea to settle the direction in the talk before building, then ask for detailed fixes one at a time, at a good pace. One early tester said that once Opus 5.5 sets the overall framework, they'd feel confident letting Sonnet 5.5 do the work of building it.
For example, the exchange might go like this.
You: I want the sign-up button to stand out a bit more. Try changing just its color. Claude: I changed the button to a deep orange that clearly stands out from the background. You: Nice. Next, shorten the description above the button to one line.
Asking for one thing at a time also makes it easier to compare what changed. Breaking it up with "just this first" instead of "fix everything at once" makes the speed count. Once the direction is set, break fixes into one at a time and stack them at a good pace.
Summary
With Sonnet 5.5, replies are faster and it's good at well-scoped work. Since you can casually ask for small fixes again and again, the loop of looking, fixing, and looking again turns that much faster.
Separating moments for careful thinking from moments for quickly moving forward on what's decided is a way of thinking you can keep using as models get newer. The flow of setting the direction in conversation and stacking fixes one at a time comes up whatever you make.
Next time you ask for a fix, try breaking it down to one thing: "just this first".