Anthropic
Claude Haiku cost & gross margin per customer
Claude Haiku is Anthropic's fast, low-cost tier. It keeps cost per customer low for high-volume features while staying capable enough for most assistant-style workloads.
Claude Haiku keeps cost low while staying capable, making it a strong default for high-frequency assistant features. Output is still several times the input price, so only very generative use cases need watching.
Input
$0.8 /Mtok
Output
$4 /Mtok
Margin per customer by usage & plan price
How Claude Haiku margin holds up as a customer's usage rises, across common subscription prices.
| Usage / mo | LLM cost | $19/mo | $29/mo | $49/mo | $79/mo |
|---|---|---|---|---|---|
| Light | $0.20 | 98.9% | 99.3% | 99.6% | 99.7% |
| Typical | $1.00 | 94.7% | 96.6% | 98% | 98.7% |
| Heavy | $4.00 | 78.9% | 86.2% | 91.8% | 94.9% |
| Power user | $16.40 | 13.7% | 43.4% | 66.5% | 79.2% |
Margin % per customer at each plan price. Token prices indicative, as of 2026-06.
Of the $1.00 a typical customer costs on Claude Haiku, output tokens are $0.60 (60%) and input $0.40. Output is priced at $4/Mtok — 5× the input rate — so the more your product generates per request, the faster a customer's margin slips.
Worked example
Take a power user on your $49/mo plan sending 8M input / 2.5M output tokens a month. On Claude Haiku that's $16.40 in tokens — that's still comfortable at 66.5% ($32.60) — even a heavy user leaves you firmly in the black on most plan prices.
How to keep Claude Haiku profitable
- Trim and cache input context — long system prompts and re-sent chat history are pure, repeated cost.
- Cap output length and stop generation early where you can: at roughly 5× the input price, every extra generated token is where Claude Haiku hurts most.
- Route easy requests to a cheaper model and reserve Claude Haiku for the hard ones that actually need it.
- Set a per-customer margin alert so one heavy user can't quietly slip into the red unnoticed.
When to choose Claude Haiku
Choose Claude Haiku when you want Anthropic quality at high frequency and low cost — chat assistants, classification, and lightweight agents you can afford to run on every plan tier.
FAQ
- How much does Claude Haiku cost per customer?
- At a typical 500k input / 150k output tokens per customer per month, Claude Haiku costs about $1.00 per customer (input 0.8/Mtok, output 4/Mtok).
- Is Claude Haiku profitable for a $49/mo AI SaaS?
- At typical usage, yes — margin is about 98% ($48.00 per customer). It erodes as usage rises; heavy and power users are where Claude Haiku can turn unprofitable.
- What's a good gross margin for an AI SaaS using Claude Haiku?
- Most AI products target a 60–80% gross margin. With Claude Haiku at typical usage you're around 98% on a $49 plan — comfortable — but your blended margin depends on the heavy users, which is the number worth watching.
- At what usage does Claude Haiku stop being profitable on a $29 plan?
- Around 14.5M input / 4.3M output tokens a month. Past that point, a $29 customer costs you more than they pay.
- How do I reduce Claude Haiku cost per customer?
- Cut output tokens first (they're the priciest), cache or trim input context, route easy requests to a cheaper model, and watch the break-even point — around 24.5M input / 7.3M output tokens a $49 customer stops being profitable.
Compare this model
Other models
Key terms