Claude Haiku 5.5

Text

1M context; lowest-cost Claude, with higher rates above 100K-token prompts

Input $0.08-$0.40 / 1M tokens | Output $0.40-$2.00 / 1M tokens Claude Haiku 5.5 API reference →

Model health Insufficient data

runapi.ai

# Base URL
https://runapi.ai

# Endpoints
POST /v1/messages

curl https://runapi.ai/v1/messages \
  -H "Authorization: Bearer $RUNAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "claude-haiku-5-5",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Review this pull request for security vulnerabilities, performance issues, and suggest concrete improvements."
    }
  ]
}'

import anthropic

client = anthropic.Anthropic(
    base_url="https://runapi.ai",
    api_key="your-runapi-key"
)

message = client.messages.create(
    model="claude-haiku-5-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Review this pull request for security vulnerabilities, performance issues, and suggest concrete improvements."}]
)

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  baseURL: "https://runapi.ai",
  apiKey: "your-runapi-key"
});

const message = await client.messages.create({
  model: "claude-haiku-5-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Review this pull request for security vulnerabilities, performance issues, and suggest concrete improvements." }]
});

https://runapi.ai 1 endpoint

15 versions available

OVERVIEW

How Claude Haiku 5.5 is billed

Anthropic released Claude Haiku 5.5 on October 7, 2026 for high-volume work such as classification, extraction, routing, summarization, and subagent tasks. It is the first Claude model priced by prompt length. When a request's prompt goes over 100,000 tokens, every rate on that request moves to a higher tier, output included. In Anthropic's list prices, each higher-tier rate is five times the short-prompt rate. The pricing table on this page lists both tiers.

RunAPI picks the tier from the whole prompt. Uncached input, cache reads, and cache writes all count toward the 100,000 tokens; output does not. A 60,000-token cached system prompt plus a 50,000-token new message adds up to 110,000 tokens, so that request pays the higher tier even though most of it is a cache read.

Haiku 5.5 uses the tokenizer introduced with Claude 4.7, which turns the same text into about 30% more tokens than Haiku 4.5 did. A prompt that stayed under 100,000 tokens on Haiku 4.5 can cross the threshold on Haiku 5.5.

On RunAPI, claude-haiku-5-5 is served through the Anthropic Messages endpoint (/v1/messages) only. RunAPI rejects requests for this model in the OpenAI Chat Completions, Responses, or Gemini formats.

Model ID

claude-haiku-5-5

Released

October 7, 2026

Context window

1M tokens

Max output

128K tokens

Higher price tier

Prompts over 100,000 tokens

Default effort (API)

medium

Protocol on RunAPI

Anthropic Messages

Playground

Try Claude Haiku 5.5

Pick any model and generate in seconds.

Specifications

Provider	Anthropic
Model ID	claude-haiku-5-5
Modality	Text
Task types	Synchronous

API endpoint	/v1/messages
Billing unit	1K tokens
Catalog status	Operational

HOW IT WORKS

Use This Model in 3 Steps

Get Your Key

Sign up for free and create an API key for claude-haiku-5-5 from the dashboard.
Send a Request

POST to /v1/messages with the claude-haiku-5-5 model slug and your parameters.
Get Results

Read the completed claude-haiku-5-5 response directly from the endpoint.

What this model offers

Endpoint coverage

Exposes 1 public endpoint(s): message.

Usage-based billing

Metered by 1K tokens with no subscription requirement.

Claude Haiku 5.5 Pricing

Endpoint

Price

message

Input: $0.08-$0.40 / 1M tokens Output: $0.40-$2.00 / 1M tokens

Prompt < 100,001 · Input: $0.08 / 1M tokens
Prompt < 100,001 · Cached input: $0.01 / 1M tokens
Prompt < 100,001 · Cache write 5m: $0.10 / 1M tokens
Prompt < 100,001 · Cache write 1h: $0.16 / 1M tokens
Prompt < 100,001 · Output: $0.40 / 1M tokens
Prompt >= 100,001 · Input: $0.40 / 1M tokens
Prompt >= 100,001 · Cached input: $0.04 / 1M tokens
Prompt >= 100,001 · Cache write 5m: $0.50 / 1M tokens
Prompt >= 100,001 · Cache write 1h: $0.80 / 1M tokens
Prompt >= 100,001 · Output: $2.00 / 1M tokens

API REFERENCE

Endpoint Specification

Endpoint: POST /v1/messages
Model ID: claude-haiku-5-5
Task Types: message

Request parameters

No request parameters are published for this endpoint.

USE CASES

Customer support

Answer customer questions from a private knowledge base, reducing ticket volume.

Document analysis

Draft contract summaries and flag key clauses for attorney review.

Code generation

Auto-generate unit tests, code reviews, and refactoring suggestions in CI.

Trusted by developers

API Calls / Day

66.7%

Success Rate

1.6s

Avg Generation

Verified customer feedback is not available for this model yet.

FAQ

Frequently asked questions about Claude Haiku 5.5

How much does Claude Haiku 5.5 cost per million tokens?

It has two sets of rates. Prompts of up to 100,000 tokens pay the lower set, and prompts over 100,000 tokens pay the higher set for input, output, cache writes, and cache reads. In Anthropic's list prices the higher set is five times the lower one. The pricing table on this page shows RunAPI's current rates for both tiers.

Which tokens count toward the 100,000-token threshold?

RunAPI counts the whole prompt of a request: uncached input tokens, cache reads, and cache writes. Output tokens do not count. The tier is decided per request, so one long request does not change the rate of the next short one.

Is Haiku 5.5 really 90% cheaper than Haiku 4.5?

On short prompts, yes. Anthropic's 90% figure compares the rates for prompts up to 100,000 tokens, and prompts over that pay 50% less than Haiku 4.5. Anthropic's own estimate of the average saving is about 75%, because the new tokenizer turns the same text into roughly 30% more tokens.

Can I use Haiku 5.5 for long agentic sessions?

It supports a 1M-token context, but once a session's prompt grows past 100,000 tokens every later request pays the higher tier. Anthropic positions Haiku 5.5 for subagents, compaction, and summarization, and recommends Sonnet 5.5 or Opus 5.5 for complex agentic coding. A Haiku subagent that gets a short, focused context stays on the lower tier.

Can I turn thinking off on Haiku 5.5?

Yes, unlike Sonnet 5.5. thinking: {type: "disabled"} is accepted at high effort or below. The API default effort is medium and adaptive thinking is on unless you disable it. Thinking tokens count toward max_tokens, so a small max_tokens can end before any text is returned.

What breaks when moving from Haiku 4.5 to Haiku 5.5?

Non-default temperature, top_p, or top_k values, assistant prefill, and thinking budget_tokens all return a 400. Forced tool_choice still works, but the response then starts with the tool call and has no thinking block. Thinking blocks are tied to the conversation, so keep the message history append-only.

Can I call Claude Haiku 5.5 with the OpenAI SDK on RunAPI?

No. RunAPI serves claude-haiku-5-5 through the Anthropic Messages API only. Use the Anthropic SDK with ANTHROPIC_BASE_URL set to https://runapi.ai and your RunAPI key.

How do I select claude-haiku-5-5?

Pass the model ID shown in the quickstart.

What's the rate limit on this variant?

Per-key rate limits scale with your usage tier. The pricing page shows current limits. If you need higher throughput, contact support to discuss tier upgrades.

Can I switch variants later?

Yes. Variant is a parameter in the request. Switch by changing the model ID — no code changes, no re-authentication, no separate billing setup. All variants share the same API key and request shape.

Does it stream?

Where streaming is available, RunAPI streams end-to-end. LLM models support token-level streaming. Media models use async task polling or webhook callbacks for result delivery.

Where do I report quality issues?

Open an issue on the public GitHub repo or email support at [email protected]. Include the task ID and model ID so the team can investigate the specific generation.

Do I need a separate provider account?

No. Your RunAPI API key is enough to access this variant and every other model in the catalog. You do not need accounts with the underlying provider.

Building with a team?

We're here to help with enterprise setup, integrations, and technical questions.