Claude Sonnet 5.5
Simon Willison’s Weblog Simon Willison ● Covered by 5 sources
Anthropic shipped Claude Sonnet 5.5, and it’s priced like Sonnet 5. It’s faster, cheaper to run, and now powers Claude’s free tier.
Based on reporting by Simon Willison’s Weblog, Simon Willison — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Anthropic has a new Sonnet model out today, and the pitch is simple: Claude Sonnet 5.5 is meant to be faster, cheaper to run, and better than Sonnet 5 without changing the sticker price. The company says it runs 30%+ faster and costs up to 30% less for most work. Simon Willison says it also appears to beat Sonnet 5 on every benchmark he looked at.
There’s a catch, because the model still has the same odd failure mode Willison saw with Opus 5.5. In “max” thinking mode, the pelican thought for 128,000 tokens, burned $1.28, and then ran out of tokens before it could finish an SVG. That is an impressive amount of effort and a very specific way to fail.
Willison also tested the higher “xhigh” thinking effort, where the model produced a pelican in 41 seconds for 5.74 cents. He says Sonnet 5.5 looks almost as good as Opus 5.5 on some coding tasks, including the viral 3D animation tricks that have been making the rounds.
The most consequential change may be where the model shows up. Sonnet 5.5 is now the model behind the free tier on claude.ai. Willison points out that OpenAI’s ChatGPT free tier uses Luna 5.6, which leaves Anthropic with the more capable free offering for now. Anthropic also says Haiku 5.5 is coming “in the coming weeks.”
My take — AI-written commentary, not fact-checked reporting
This is the sort of move that actually matters: better free-tier software beats a lot of loud product marketing. The pelican benchmark circus is funny, but the real story is that Anthropic is using a stronger model to make the free tier look less like a sampler and more like a threat. That’s the kind of competition users should want, even if the birds keep cost accounting weird.
Read more about this at: Simon Willison’s Weblog
