claude-haiku-4-5-20251001

Claude Haiku 4.5 is a model in the Claude family that emphasizes speed and high throughput, and is suitable for production applications that are sensitive to latency and have a large number of calls. It can be used for real-time customer service, intent recognition, text classification, information extraction, short summaries, content security preliminary screening, batch document processing and workflow routing. Compared with the large Opus model, it is more suitable for completing steps with clear structure and quick verification, and can also be used as a pre-screening layer in a multi-model architecture. It supports manually extended thinking that can be adjusted between speed and quality for complex tasks.

Context

200000 tokens

Max output

64000 tokens

Input

llm image

Output

llm