Anthropic · Workhorse

Claude Sonnet 5.5 API pricing and specs

Anthropic's mid-priced model, with five breaking changes from Sonnet 5, including thinking that can be limited to the gaps between tool calls.

Input

⁨$2⁩

Cached input

⁨$0.20⁩

Output

⁨$10⁩

US dollars per million tokens, standard tier, short-prompt rate

Default effort is high, which spends more output tokens than Opus 5.5's medium default; lower it for routine steps. Cache writes cost $2.50 per million for 5 minutes.

Spec sheet

API model string: claude-sonnet-5-5
Provider: Anthropic
Context: 1M tokens
Max output: 128K tokens
Released: September 28, 2026
Batch API: 50% off, results within hours
Documented agent features: Tool callingComputer useAdjustable reasoning

What a task costs

Four reference workloads priced at this model's list rates, with caching applied after the first call. Your numbers will differ; the calculator lets you change every assumption.

Workload	Per task	Per 1,000 tasks
Support bot 4 calls, short tool lookups	⁨$0.056⁩	⁨$56⁩
Coding agent 30 calls, files and test output	⁨$1.98⁩	⁨$1,977⁩
Browser agent 25 calls, screenshots and page text	⁨$1.40⁩	⁨$1,397⁩
Research agent 15 calls, long search results	⁨$1.32⁩	⁨$1,323⁩