Skip to main content

Pricing · preview

Pay for what the model reads

Needle never generates text, so there are no output tokens to bill. You pay for input tokens only: the state plus each question.

Free

$0 10M input tokens / month

About 20,000 typical requests. No card required.

  • needle-1
  • All three primitives
  • Rate limited
Get access

Pay as you go

$0.20 per 1M input tokens

About $0.10 per 1,000 typical requests. Output is always $0.

  • Everything in Free
  • Higher rate limits
  • Email support
Get access

Dedicated

Custom from $1,500 / month

A model tuned to your policies and wording, on reserved capacity.

  • Tuned on your decisions
  • Reserved throughput
  • Evaluation on your holdout set
  • Direct line to the lab
Talk to us

How tokens are counted

A typical request is a few hundred tokens of state and three questions, around 500 input tokens in total. The usage field on every response reports exactly what was counted.

Preview pricing may change before general availability. We will email everyone on the list before it does.