claude-haiku-4-5 API

Claude Haiku 4.5 sits at the small-and-fast end of the Claude family: it trades a slice of peak reasoning depth for dramatically lower latency and a 5x lower price than Sonnet-tier models. Anthropic positions it as "the fastest model with near-frontier intelligence", and in practice it lands close to previous-generation mid-tier models on everyday tasks while responding noticeably quicker. If your workload is high-volume and latency-sensitive rather than research-grade, this is usually the right Claude to start with. The model carries a 200K-token context window with up to 64K output tokens per request — enough to feed in long documents, full support-ticket histories, or several source files at once, while the per-request cost stays small. Typical sweet spots are classification and routing, structured extraction, summarization, autocomplete, and real-time chat. It also works well as the "worker" model in multi-agent setups: a stronger model plans, and Haiku 4.5 fans out to execute many cheap sub-tasks in parallel. On CaMeL Hub, claude-haiku-4-5 is exposed through both the OpenAI-compatible endpoint (/v1/chat/completions — point any OpenAI SDK at https://api.camel-hub.com/v1 and change only the base_url and key) and the Anthropic-native Messages format, with streaming supported on both. Billing is metered per token with separate input and output rates, you can pay by top-up or subscription plan, and Alipay, WeChat Pay, and international cards are all accepted.

curl https://api.camel-hub.com/v1/chat/completions \
  -H "Authorization: Bearer $CAMEL_HUB_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-haiku-4-5",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from openai import OpenAI

client = OpenAI(
    api_key="$CAMEL_HUB_KEY",
    base_url="https://api.camel-hub.com/v1",  # <-- the only change vs. the OpenAI default
)

resp = client.chat.completions.create(
    model="claude-haiku-4-5",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.CAMEL_HUB_KEY,
  baseURL: "https://api.camel-hub.com/v1", // <-- the only change vs. the OpenAI default
});

const resp = await client.chat.completions.create({
  model: "claude-haiku-4-5",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(resp.choices[0].message.content);