Anthropic · Small and fast

Claude Haiku 4.5 API pricing and specs

The only small Claude model still offered, with vision and manually budgeted extended thinking.

Input

⁨$1⁩

Cached input

⁨$0.10⁩

Output

⁨$5⁩

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

Anthropic says it will be retired no sooner than 15 October 2026, so plan a replacement before you build on it.

Spec sheet

API model string: claude-haiku-4-5-20251001
Provider: Anthropic
Context: 200K tokens
Max output: 64K tokens
Released: —
Batch API: 50% off, results within hours
Documented agent features: Tool callingAdjustable reasoningImage input

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	⁨$0.028⁩	⁨$28⁩
Coding agent 30 calls, files and test output	⁨$0.989⁩	⁨$989⁩
Browser agent 25 calls, screenshots and page text	⁨$0.698⁩	⁨$698⁩
Research agent 15 calls, long search results	⁨$0.662⁩	⁨$662⁩