Anthropic
Claude Haiku 4.5
Fast, low-cost Claude model for high-volume and latency-sensitive tasks.
USD per 1M output tokens
$5.00
- input
- $1.00
- output
- $5.00
#12 cheapest of 25 in Large language models
Official pricing ↗Capabilities, explained
What this API can do, and what each capability means for your project.
Input price
$1
Cost of the text, images and files you send to the model.
Context window
200K tokens
How much text the model can read at once. 200K tokens ≈ a 500-page book.
Max output
64K tokens
Longest single response the model can write.
Input types
text, image, pdf
What kinds of input the model understands besides text.
Tool use / function calling
Yes
The model can request calls to your functions/APIs with structured arguments, the basis of agents.
Guaranteed JSON output
Yes
The API can force the response to match a JSON schema you provide.
Extended reasoning
Yes
The model can 'think' step by step before answering, improving hard math/coding at extra cost and latency.
Open weights
No
The model weights are downloadable, so you can self-host it or switch hosting providers without changing models.
Typical latency
400 ms
How long until the first useful output arrives. Lower is better for interactive apps; batch jobs rarely care.
Price history
Only one price recorded so far. The chart appears once the price changes.