One subscription. Every open model.
Chat, Koda and the API on one plan. Limits refill on rolling windows, so a busy afternoon never costs you the week.
Light coding, an hour here and there.
- Per 5 hours
- 500
- Per week
- 2,400
- Per month
- 5,000
- 11 models hosted on 0G
- Private models in a secure enclave
- No per-minute limit
A couple of hours of coding.
- Per 5 hours
- 700
- Per week
- 2,800
- Per month
- 10,000
- Every model in the catalog
- Image understanding and web search
Most of the working day.
- Per 5 hours
- 1,500
- Per week
- 6,500
- Per month
- 18,400
- Every model, higher per-minute limits
- Image understanding and web search
All day, agents running.
- Per 5 hours
- 3,000
- Per week
- 15,000
- Per month
- 45,000
- Every model, highest per-minute limits
- Image understanding and web search
Try it before you pay. Chat, Koda and the API on Freemium, Lite and Pro tier models, no card needed.
- Per day
- 50
- Per month
- 200
Prices in USD, billed monthly. Change plans or cancel any time. A request is one standard model call; lighter calls spend less, flagships more.
Every plan, side by side.
| Feature | Free | Go | Lite | Pro | Max |
|---|---|---|---|---|---|
| How much · request credits | |||||
| Price a month | $0 | $6.99 | $14.99 | $29 | $79 |
| Per day | 50 | — | — | — | — |
| Per 5 hours | — | 500 | 700 | 1,500 | 3,000 |
| Per week | — | 2,400 | 2,800 | 6,500 | 15,000 |
| Per month | 200 | 5,000 | 10,000 | 18,400 | 45,000 |
| How fast · requests a minute, per key | |||||
| Freemium-tier models | 30 | no limit | 60 | 120 | 200 |
| Lite-tier models | 5 | no limit | 10 | 30 | 120 |
| Pro-tier models | 5 | no limit | 8 | 15 | 60 |
| Max-tier models | — | no limit | 4 | 8 | 40 |
| Models | |||||
| Model tiers you reach | Freemium, Lite, Pro | All four, on 0G | All four | All four | All four |
| Included | |||||
| Chat, Koda and the API | |||||
| Web search, scrape and OCR tools | |||||
| Vector DB storage per key | 20 MB | 20 MB | 200 MB | 800 MB | 2 GB |
| Bring your own keys | |||||
The windows roll: each request ages out of the 5-hour, weekly and monthly windows on its own, with no calendar reset.
What a request costs.
Each call spends request credits by the model's weight. Most coding calls are standard, so the limits above are close to the number of calls you can make.
Short, and meant to be accurate.
Anything else, email support@openadapter.dev or ask on Discord.
Start free →One call to a standard model spends one request credit. Light models, embeddings and web tools spend less, flagships two to three and the top tier five, so most coding calls are close to one each.
They roll. Each request ages out of the 5-hour, weekly and monthly windows on its own, so there is no reset day to wait for. Your dashboard shows what is left in each.
Calls pause with a 429 until older requests age out of that window. Going faster than your per-minute pace only slows you for a few seconds. You can move to a bigger plan any time.
No. Add keys from OpenCode, Z.AI Coding Plan, Anthropic, OpenAI and others in Settings; those requests use none of your plan.
You keep your plan until the end of the period you paid for, then move to Free. If your plan was cheaper when you joined, that price stays yours while you are subscribed; cancel and come back, and you pay the price of the day.
Paid plans are non-refundable unless the law where you live says otherwise. Payments are handled by DodoPayments and Paddle. The full policy is on the refund page.
