Needle · planning assumptions
Proposed pricing
The inference API is in development and is not yet available. This cost model is a proposal, not a purchasable plan or a price commitment.
Estimator assumptions
- Proposed monthly allowance
- 10,000,000 input tokens
- Above that allowance
- $0.20 per million input tokens
- Output tokens
- None generated by the model
Input tokens include the supplied state and each question. Hosted limits, support, dedicated capacity, and any tuning offering have not been committed.
Illustrative estimate
Explore the cost model
- Input tokens
- 173.3M
- Proposed allowance
- 10M
- Estimated monthly
- $33
Assumes 40 tokens per question. This is a planning estimate, not a bill or a quote.
How tokens are counted
The estimator divides the number of decisions by the questions per request, rounding up to whole requests. It counts the state once per request and assumes 40 tokens per question. Actual tokenization and hosted billing are not implemented here.
These assumptions may change before an API is available. Follow development or discuss your requirements.