Claude Haiku 4.5 token cost & pricing
Claude Haiku 4.5 is Anthropic's fast, low-cost tier at $1 per million input and $5 per million output, with the same 200K context window and 64K max output as its larger siblings. That combination — small model, big context — is unusual and makes it useful for long-document work that does not need frontier reasoning.
Claude Haiku 4.5 specifications
| Vendor | Anthropic |
|---|---|
| Input price | $1 / 1M tokens |
| Output price | $5 / 1M tokens |
| Cached input price | $0.1 / 1M tokens (90% cheaper) |
| Context window | 200,000 tokens |
| Max output | 64,000 tokens |
| Open weights | No |
| Price rank | 9 of 17 (cheapest first) |
| Best for | budget, fast |
What Claude Haiku 4.5 costs in practice
Four common workloads, priced at list rates. Monthly figures assume 30 days.
| Workload | Per request | Requests/day | Per month |
|---|---|---|---|
| Chatbot · 1K in / 500 out | $0.0035 | 10,000 | $1,050 |
| RAG Q&A · 8K in / 400 out | $0.01 | 5,000 | $1,500 |
| Doc summary · 30K in / 1K out | $0.03 | 1,000 | $1,050 |
| Code agent · 60K in / 3K out | $0.07 | 500 | $1,125 |
How many requests does $100 buy?
Assuming a 1,000-token prompt and a 500-token response per request.
How Claude Haiku 4.5 pricing actually works
Roughly a third of Sonnet's price on both axes, with cached input at $0.10 per million. The 5x input-to-output ratio means generation-heavy workloads pay proportionally more, so keep answers concise where the product allows it.
When to use Claude Haiku 4.5
Use it for support chat, summarization, extraction from long documents, and as the cheap tier in a routed pipeline. It is a reasonable default for anything that must feel instant — latency is its main practical advantage over Sonnet.
How to cut your Claude Haiku 4.5 bill
Cache your system prompt and any reference documents, then keep per-request additions small. Because the context window is 200K, there is a temptation to dump entire corpora into each call — resist it, since even cheap input tokens add up at volume.
Other Anthropic models
Cheaper alternatives to Claude Haiku 4.5
- Kimi K2Moonshot48% cheaper
- Gemini 2.5 FlashGoogle53% cheaper
- DeepSeek R1 (reasoner)DeepSeek54% cheaper
- GPT-5 miniOpenAI63% cheaper
Run the numbers on your own prompt
Paste your actual text into the token counter to get an exact count instead of an estimate. Nothing is uploaded.
Open the calculator →