claude-haiku-5.5 API — per-token pricing and code examples
Price per 1M tokens
Cache rates apply when the provider reports cache tokens.
68% offAbout this model
claude-haiku-5.5 is served through OpenAI Chat Completions, OpenAI Responses, Anthropic Messages. You pay $0.032 per 1M input tokens and $0.16 per 1M output tokens. That is 68% below the official price of $0.10 and $0.50. Cache reads cost $0.0032 per 1M tokens when the provider reports cached tokens. It is included in budget packs from $1.99 up.
Where Claude Haiku 5.5 fits
Claude Haiku 5.5 is Anthropic's newest small model, released on October 7, 2026. Anthropic describes it as the model "for high-volume, latency-sensitive tasks such as classification, extraction, and routing". Retirement is listed as not sooner than October 7, 2027, and its knowledge reaches June 2026. It reads text and images, answers in text, and supports 1M tokens of context with up to 128K output tokens in one synchronous reply.
Pick Haiku 5.5 for work that runs many times with short answers: tagging tickets, pulling fields out of documents, choosing which larger model should take a request, or the background tasks of a coding agent such as naming a session. For long reasoning over a large codebase, a Sonnet or Opus model is the better fit.
Thinking is on unless you turn it off
Unlike the 4.6 models, Haiku 5.5 thinks by default. A request with no thinking field runs with adaptive thinking, and the default effort is medium. All five effort levels are accepted: low, medium, high, xhigh and max.
Three rules from Anthropic's thinking guide decide what you can send:
thinking: {type: "disabled"}turns thinking off, but only athigheffort or below. Withxhighormaxit returns a 400 error.- The older
enabledmode withbudget_tokensis rejected with 400 on this model. Use adaptive. - The default
displayisomitted, so the reasoning text is hidden. Setdisplay: "summarized"if you want to read a summary.
Anthropic also notes that with thinking off, the model can skip a tool call it needs when you also ask for JSON output. If a workflow depends on a tool being called, keep adaptive thinking on at low effort instead of disabling it.
What it costs on Cent AI
Haiku 5.5 is charged per token from your prepaid USD balance, at the rates in the price box on this page. Because thinking is on by default, reasoning tokens are counted as output even when the visible answer is one word. For classification and routing, set output_config.effort to low, or disable thinking, to keep each answer small.
The minimum cacheable prompt on Haiku 5.5 is 512 tokens, lower than on the 4.x models, so even a short system prompt can be cached. On Cent AI, cache writes for the 5-minute and the 1-hour lifetime each have their own rate in the price box, and cache tokens are billed only from the counts the model reports.
Using it from your tools
Haiku 5.5 answers on /v1/messages, /v1/chat/completions and /v1/responses. Cent AI does not change your parameters silently, so Anthropic's rules for the model apply. Anthropic's guidance is to omit temperature, top_p and top_k. In Claude Code, ANTHROPIC_DEFAULT_HAIKU_MODEL=claude-haiku-5.5 points the background tasks at this model.
A /v1/messages request for a quick classification, with thinking off at low effort:
{
"model": "claude-haiku-5.5",
"max_tokens": 64,
"thinking": { "type": "disabled" },
"output_config": { "effort": "low" },
"messages": [{ "role": "user", "content": "Label this ticket: billing, bug or question" }]
}
Example: the cost of one request
Take a request that sends 2,000 new input tokens, reads 8,000 tokens from cache and gets 800 output tokens back. At the rates in the price box on this page:
- input: 2,000 × $0.032 / 1,000,000 = $0.000064
- cache read: 8,000 × $0.0032 / 1,000,000 = $0.0000256
- output: 800 × $0.16 / 1,000,000 = $0.000128
That request costs $0.0002176, about Rp4 at Cent AI's exchange rate on 2026-11-15 (Rp18,229.36 per USD). A thousand requests like it cost $0.2176. Your real requests will differ; each one shows its own amount in usage.cost and on the Requests page.
Questions
Is there a 5-hour or weekly limit? No. You pay per token from your balance. Each key accepts up to 600 requests per minute, and you can set your own daily or monthly limit on a key.
Is a failed request charged? A request that fails before the model receives it costs $0. Tokens the model already produced are charged.
Does unused balance expire? No, not while your account exists. Every top-up is manual, with no auto-renewal.
Related models
Compare with claude-sonnet-5.5, gpt-6-luna, deepseek-v4.1-flash, or see every model on the pricing page.
Specification
- Context window
- —
- Input and output
- Text
- Endpoints
- /v1/chat/completions, /v1/responses, /v1/messages
- Smallest pack
- $1.99
Latency and error rate per model are on the status page. status.cent-ai.dev
What a request costs
Open the full calculatorCost per request
Estimate
claude-haiku-5.5An estimate at price version 4. Your bill comes from the usage the provider reports for each request.
Save $0.0161Code
curl https://api.cent-ai.dev/v1/chat/completions \
-H "Authorization: Bearer $CENT_AI_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "claude-haiku-5.5", "messages": [{"role": "user", "content": "Hello"}]}'Set CENT_AI_KEY to a key from the dashboard. OpenAI SDKs use https://api.cent-ai.dev/v1; Anthropic SDKs use https://api.cent-ai.dev.
Price history
Version 4, in effect since Nov 14, 2026.
Effective
per 1M tokens
Inputold $0.02562 new $0.032Cache readold $0.002562 new $0.0032Cache writeold $0.03203 new $0.04Cache write 1hold $0.05124 new $0.064Outputold $0.1281 new $0.16Effective
per 1M tokens
Inputold $0.027 new $0.02562Cache readold $0.0027 new $0.002562Cache writeold $0.027 new $0.03203Cache write 1hold $0.027 new $0.05124Outputold $0.135 new $0.1281Effective
per 1M tokens
Inputold $0.027 new $0.027Cache readold $0.0027 new $0.0027Cache writeold $0.03375 new $0.027Cache write 1hold $0.054 new $0.027Outputold $0.135 new $0.135
Compare with
- deepseek-v4.1-flashInput$0.015
- gpt-6-lunaInput$0.01978
- qwen3.8-flashInput$0.015
- deepseek-v4-flashInput$0.015
Questions
How much does claude-haiku-5.5 cost on Cent AI?
$0.032 per 1M input tokens and $0.16 per 1M output tokens. Cache reads cost $0.0032 and cache writes $0.04 per 1M tokens. Each request is rounded up once.
Is there a minimum charge?
A successful request costs at least $0.00002. A request that ends early, because the client left or the output was cut off, pays only for its tokens.
How do I pay?
Top up a prepaid USD balance with QRIS or USDT/USDC, or buy a budget pack once. Nothing renews automatically.
What if the provider does not report usage?
The request is billed from an estimate and marked estimated on the Requests page. An estimate never contains cache tokens.
Which endpoint do I call?
/v1/chat/completions, /v1/responses, /v1/messages. The OpenAI base URL is https://api.cent-ai.dev/v1.
Try claude-haiku-5.5 with one key
One prepaid balance for every model. Pay with QRIS or USDT/USDC.