Pricing
TAO-1 is a frontier-level AI system powered by Bittensor, billed per token on what you actually use: no subscriptions, no seat minimums. The OpenAI-compatible API meters every call in real time.
Updated
New accounts start with $5 of free credit. No payment setup needed.
How much does TAO-1 cost?
Price per million tokens
| Model | Input | Cached input | Output |
|---|---|---|---|
| TAO-1 balanced | $0.50 | $0.15 | $7.50 |
| TAO-1 Fast our fastest, lowest-cost dial | $0.25 | $0.075 | $4.00 |
| TAO-1 Best coming soon | Price to be announced | ||
Prices are per million tokens, charged on the tokens TAO-1 reports for each call. For comparison, Moonshot AI publishes Kimi K3 at $3.00 input, $0.30 cached input and $15.00 output per million tokens (platform.kimi.ai/docs/pricing/chat, checked October 14, 2026): TAO-1's input price is one sixth of that and its output price half; TAO-1 Fast's is one twelfth and just over a quarter.
Is TAO-1 right for you?
A good fit if you want to pay only for what you use, with no monthly fee. TAO-1 Fast is the lower-cost choice when price or speed matters most.
Less of a fit if you need a fixed monthly bill: your cost depends on how many tokens you send and receive.
How does billing work?
Every non-streaming /v1/chat/completions response carries the tokens it was billed on, and the charge for that call is in a response header — see the API reference. For a streaming response, add stream_options: {"include_usage": true} to receive the billed tokens in a final chunk. Cached input is whatever the model reports as reused from a prior turn; everything else is billed at the input rate.
A new account starts with $5 of free credit so you can make your first calls right away. After that, add credit from your account page.
Questions about billing
- Is there a monthly fee?
- No. You pay only for the tokens you send and receive.
- Does TAO-1 charge for failed requests?
- No. A request that errors before TAO-1 returns usage is not charged.
- What happens if a stream is cut off early?
- You're billed for the tokens actually generated before the cut-off, not a flat rate.