The sticker price did not move from Sonnet 5 to Sonnet 5.5, so "how much does Sonnet 5.5 cost" really has two answers: the per-token list price, which is the same, and the cost of finishing a task, which Anthropic says drops by up to 30% because the new model writes shorter outputs and makes fewer tool calls. For agent loops and coding tools, the second number is the one on your invoice.
All figures are Anthropic’s published list prices per million tokens. Opus 5.5 is shown for comparison because most people pricing Sonnet 5.5 are deciding between the two.
| Per 1M tokens (list) | Sonnet 5.5 | Sonnet 5 | Opus 5.5 |
|---|---|---|---|
| Input | $2 | $2 | $4 |
| Output | $10 | $10 | $20 |
| Cache read | $0.20 | — | $0.20 |
| Cache write | $2.50 | — | $5 |
| On apimodels.app | not yet listed | $1.60 / $8 | $2.40 / $12 |
For a typical coding turn that sends 50,000 input tokens (40,000 of them cached) and returns 3,000 output tokens, Sonnet 5.5 at list price costs about $0.008 for cached input, $0.02 for fresh input and $0.03 for output: roughly $0.058 per turn. The same turn on Opus 5.5 is about twice that. Because Sonnet 5.5 tends to finish in fewer turns and tool calls, the per-task gap to Sonnet 5 is where the savings show up, not in the price table.
Cache writes cost 1.25x input and cache reads a tenth of input, so long system prompts and repository context should be cached; tools like Claude Code do this automatically.
cURL
curl https://api.apimodels.app/v1/messages \
-H "x-api-key: $APIMODELS_API_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{"model": "claude-sonnet-5", "max_tokens": 16000,
"messages": [{"role": "user", "content": "Summarise the risks in this pull request: ..."}]}'$2 per million input tokens and $10 per million output tokens on the Anthropic API, with cache reads at $0.20 and cache writes at $2.50 per million. That is unchanged from Sonnet 5 and half the price of Opus 5.5.
Per token, no: the prices are identical. Per task, Anthropic reports up to 30% lower cost because Sonnet 5.5 uses fewer tokens and tool calls and generates output more than 30% faster.
Not yet. Anthropic said on 22 September 2026 that Haiku 5.5 will follow Opus 5.5 in the coming weeks, but has not published a release date, price or benchmarks. Until then, Sonnet 5.5 is the cheapest Claude 5.5 model.
Not yet. Sonnet 5 is available now at $1.60 / $8 per million tokens, 20% below list, and Opus 5.5 at $2.40 / $12, on the same key. This page will carry the Sonnet 5.5 price the day it is added.
