Claude Haiku 5.5 API pricing
Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarisation, extraction, sub-agents and browser use, with a 1M-token context window. On FastInfra the Claude Haiku 5.5 API costs $0.11 per million input tokens and $0.54 per million output tokens, billed per token through an OpenAI-compatible API.
Prices synced November 3, 2026 · Model ID anthropic/claude-haiku-5.5
Claude Haiku 5.5 specs and pricing
Claude Haiku 5.5 API pricing per token
You pay for prompt tokens at the input rate and generated tokens at the output rate. No subscription, seat fee or minimum — credit never expires.
| Rate | Input / 1M tokens | Output / 1M tokens | Input / 1K tokens |
|---|---|---|---|
| Standard (real-time) | $0.11 | $0.54 | $0.000108 |
Batch (async) · anthropic/claude-haiku-5.5:batch |
$0.05 | $0.27 | $0.000054 |
What Claude Haiku 5.5 costs in practice
Worked examples at the standard rate above. Real token counts depend on your prompts.
| Workload | Tokens | Cost |
|---|---|---|
| 1,000 chatbot replies | 1,500 input + 400 output tokens each | $0.38 |
| Summarise a 100-page report | ≈60,000 input + 1,500 output tokens | $0.0073 |
| RAG answers over your docs, 10,000 queries | 4,000 input + 300 output tokens each | $5.94 |
| Coding-agent session | ≈250,000 input + 25,000 output tokens | $0.04 |
| 1M input + 1M output tokens | the per-million list rates combined | $0.65 |
Claude Haiku 5.5 cost calculator
Enter your traffic to estimate daily and monthly spend at the live FastInfra rate.
When to use Claude Haiku 5.5
Haiku 5.5 succeeds Claude Haiku 4.5 with stronger coding, computer use and tool use, while staying in the price band where you can run millions of requests a day. It is the Claude model to put behind classification pipelines, ticket triage, chat summaries, retrieval-augmented answers and the worker agents that a larger planner model delegates to.
It keeps the 5-series features — image and file input, tool calling, structured JSON output and adjustable reasoning — so moving a prompt from Sonnet 5 to Haiku 5.5 is usually a model-string change plus an evaluation run, not a rewrite.
Good fit
- Summarisation, tagging and data extraction at scale
- Sub-agents and tool-calling workers under a larger planner
- Latency-sensitive chat where answers are short
- RAG answers over your own documents
Consider another model if
- Deep multi-step reasoning or large refactors — use Sonnet 5 or Opus 5
- Tasks where one wrong answer is very costly
- You need the very lowest price — DeepSeek V4 Flash and Gemini Flash-Lite tiers are cheaper still
How to call the Claude Haiku 5.5 API
Works with the official OpenAI SDKs — change the base URL and API key only. One key covers every model on FastInfra.
| Field | Value |
|---|---|
| Model ID | anthropic/claude-haiku-5.5 |
| Capability | Chat (vision) |
| Base URL | https://api.fastinfra.ai/v1 |
| Primary endpoint | POST https://api.fastinfra.ai/v1/chat/completions |
| Auth | Authorization: Bearer YOUR_API_KEY |
| OpenAI Chat SDK | Yes — set base_url to the Base URL |
| Validate model id | GET https://api.fastinfra.ai/v1/models/anthropic/claude-haiku-5.5 |
| Docs | Integration reference |
Use multimodal message content (image_url). Same base URL as text chat.
Python
from openai import OpenAI
client = OpenAI(api_key="YOUR_API_KEY", base_url="https://api.fastinfra.ai/v1")
stream = client.chat.completions.create(
model="anthropic/claude-haiku-5.5",
messages=[{"role": "user", "content": "Explain vector databases in two sentences."}],
stream=True,
)
for chunk in stream:
print(chunk.choices[0].delta.content or "", end="")Node.js
import OpenAI from "openai";
const client = new OpenAI({ apiKey: process.env.FASTINFRA_API_KEY, baseURL: "https://api.fastinfra.ai/v1" });
const response = await client.chat.completions.create({
model: "anthropic/claude-haiku-5.5",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(response.choices[0].message.content);curl
curl https://api.fastinfra.ai/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "anthropic/claude-haiku-5.5",
"messages": [{"role": "user", "content": "Hello!"}]
}'Served over 1 independent route. Requests go to the lowest-cost healthy route and fail over automatically if one is unavailable. Live latency and uptime are on the status page.
Claude Haiku 5.5 vs alternatives
Live FastInfra prices per 1M tokens. Every model uses the same API key, so switching is a one-line change.
| Model | Input | Output | Context | |
|---|---|---|---|---|
| Claude Haiku 5.5 (this page) | $0.11 | $0.54 | 1M | |
| Claude Sonnet 5 | $2.16 | $10.80 | 1M | Compare |
| Gemini 3.8 Flash | $0.81 | $4.05 | 1M | Compare |
| DeepSeek V4 Flash | $0.10 | $0.20 | 1M | Compare |
| GPT-6 Luna | $0.11 | $0.54 | 1.05M | Compare |
| GPT-OSS 120B | $0.16 | $0.65 | 128K | Compare |
| Claude Haiku 4.5 | $1.08 | $5.40 | 200K | Compare |
| Claude Sonnet 5.5 | $2.16 | $10.80 | 1M | Compare |
Claude Haiku 5.5 API — frequently asked questions
How much does the Claude Haiku 5.5 API cost?
On FastInfra, Claude Haiku 5.5 (anthropic/claude-haiku-5.5) costs $0.108/1M input, $0.54/1M output tokens. Billing is per token used from prepaid credit, with no subscription or minimum spend.
What is the Claude Haiku 5.5 context window?
Claude Haiku 5.5 accepts up to 1,000,000 tokens of context (about 1M) and can generate up to 128,000 output tokens per response. Prompt and response tokens share that window.
Is Claude Haiku 5.5 compatible with the OpenAI SDK?
Yes. Point the OpenAI Python or Node SDK at https://api.fastinfra.ai/v1, use a FastInfra API key and set model to "anthropic/claude-haiku-5.5". Chat completions, streaming, tool calling and JSON structured output use the same request shape.
Does Claude Haiku 5.5 support tool calling and vision?
Claude Haiku 5.5 supports tool (function) calling, JSON structured output, configurable reasoning and image input.
Is Claude Haiku 5.5 cheaper than Haiku 4.5?
Yes. Compare both on FastInfra: Haiku 5.5 is listed at a lower per-token rate than Claude Haiku 4.5 while Anthropic reports stronger coding and tool use.
What is Claude Haiku 5.5 good for?
High-volume, cost-sensitive work: summarisation, extraction, classification, sub-agents that call tools, and browser or computer-use automation where many short steps add up.
Start using Claude Haiku 5.5 today
Create a free account, add credit from $5 and make your first Claude Haiku 5.5 request in under five minutes. Compare every model on the pricing page.