Announced 28 Sept 2026 · Sources checked
Is Sonnet 5.5 cheaper per token?
The launch page lists the same rates as Sonnet 5: $2 per million input tokens and $10 per million output tokens. The “up to 30% less” claim concerns task cost.
What should you measure besides output speed?
Faster text generation does not account for tool calls, retries, or review.
| Measure | What it tells you |
|---|---|
| Time to first useful response | How quickly the user gets something actionable |
| Time to finished task | How long the whole workflow takes |
| Human correction time | How much work remains after generation |
How can you test the difference?
Choose a repeatable task and keep the conditions consistent.
- Use the same instructions, tools, and review standard.
- Record usage and total time for successful and failed attempts.
- Keep the settings with your results so you can repeat the comparison.
Common questions
Will every task cost 30% less?
No. Anthropic reports up to 30% lower task cost in its testing; results depend on the task and settings.
Has AiLookout benchmarked it?
No. The launch measurements are attributed to Anthropic.
What to remember
Anthropic’s cost claim concerns the work needed to complete a task; it does not mean every token is 30% cheaper.
Sources & further reading
How this story was made
Written by Kristian Kostov with AI assistance and checked against the linked sources. Company performance claims are attributed to the company. Analysis reflects AiLookout’s interpretation; we have not independently tested the products discussed. Cover photography is illustrative and does not depict the specific announcement or product.
Our editorial standards
