Claude Sonnet 5.5 ships faster and cheaper, debuts #2 on Vals Index
- Source
- x.com
- Date
Claude Sonnet 5.5 is now available: https://t.co/DuxGkiPLRY
Context
Anthropic's claim that Claude Sonnet 5.5 costs less refers to the work completed, rather than a reduction in its published rates. The company kept Sonnet 5's prices of $2 per million input tokens and $10 per million output tokens, but says the newer model typically needs far fewer tokens for the same work and, in its testing, costs up to 30% less per task than Sonnet 5. It also reports output generation more than 30% faster. The release explanation supplies the detail missing from the short announcement.
For a metered application, fewer tokens can lower the bill even when the rate stays fixed. The size of that benefit depends on the work and the effort setting: Anthropic says lower effort levels make Claude answer faster and use fewer tokens, which suits routine work. The savings figure is the company's testing result, not a guarantee for every workload. Cost and speed details.
- token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
Checking sign-in…
Loading comments…





