Use it through the API.
Call eligible models during the serving window. Usage draws down your $20-per-OSTR credit at the series’ fixed model prices.
Forward Market · OSTR
Monthly inference.
Yours to use or trade.
OSTR is a tradable right to AI inference for a specific month. Use it across eligible models, or offer unused OSTR for sale while trading is open.
Live Series
Orders execute when a buyer and seller match.
How it works
Use it.
Trade it.
Buy OSTR for the month you need. Once you hold OSTR, choose how much to use and how much to trade.
You can use part of your balance and offer the rest for sale. You don’t have to use inference before trading.
Buy OSTR
Choose a serving month and place an order on the market.
Trade it on the market.
Offer unused, uncommitted OSTR for sale while trading is open. Your order fills when a buyer matches your price.
The model menu
One balance.
A menu of models.
Each OSTR provides $20 of inference credit across the eligible models in your series. Choose where to use it; you don’t have to split usage evenly across the menu.
The $20 face value and series model prices are fixed. The market price you pay for OSTR can change.
- Models
Choose an eligible model.
Use the eligible models in your series’ menu. Official model prices are fixed when published; new models may be added during the serving month.
- Face value
$20 of inference credit per OSTR.
Call eligible models through the API. Usage draws from your $20-per-OSTR credit at the series’ listed prices.
- Savings
Your purchase price determines your savings.
Buying one OSTR for $16 gives you $20 of inference credit: 20% savings relative to the series’ prices when you use the full credit.
See a usage and savings example
Example: run GLM 5.3 for 20% less
Suppose you buy 1 OSTR for $16. It gives you $20 of inference credit.
Your series lists GLM 5.3 at $1.40 per million input tokens and $4.40 per million output tokens.
| Usage | Credit used |
|---|---|
| 8 million input tokens | $11.20 |
| 2 million output tokens | $8.80 |
| Total | $20.00 |
You’ve used your 1 OSTR in full. You paid $16 for $20 of inference — saving $4, or 20%, against the series’ listed prices.
Example purchase price. Use your credit within the series’ serving month.
Terms & FAQ
Before you trade.
Check the serving dates, model menu and delivery terms for your series before placing an order.
Check the series’ terms for model changes and delivery conditions.
Series terms at a glance
- Serving month
- Use your credit between the series’ stated start and end times, in UTC.
- Models & pricing
- Official model prices are fixed when published in the series’ menu. New models may be added during the serving month.
- Face value
- Each OSTR provides $20 of inference credit.
- Trading
- Sell unused OSTR before the serving period ends. A matching buyer is required.
- Expiry
- Unused credit expires. No automatic rollover or cash redemption.
See your series’ published terms for full delivery conditions. See fees
When can I start using my OSTR?
From your series’ stated start time until its end time, in UTC.
Can I sell OSTR without using it first?
Yes. Sell unused OSTR before the serving period ends. Your order fills when a buyer matches your price.
What do OW and CW mean?
OW means open-weight; CW means closed-weight. Use the eligible models listed in your series’ menu.
What happens to unused OSTR?
Unused credit expires at the end of the serving period. No automatic rollover or cash redemption.
What if a model changes or a provider cannot deliver?
Model changes and delivery remedies follow your series’ published terms. Switching to a different model requires your consent.
How does OSTR fit with pay-as-you-go?
Buy OSTR for planned monthly usage. Each OSTR provides $20 of inference credit at your series’ fixed model prices. Use pay-as-you-go for variable demand or models outside your menu. Explore pay-as-you-go models.
What fees apply?
No additional API fee or OSTR delivery commission. Maker and taker fees are waived during launch. See fees.
Ready to trade OSTR?
Choose your series and place an order on the market.
Want to supply inference or make markets? Talk to us