OpenScale / Pricing
Pricing
Pay per token with no minimum, or pay one flat monthly price for coding agents. New accounts get free credit to try it.
Pay as you go, per 1M tokens
| Model | Input | Cached input | Output |
|---|---|---|---|
| DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash | $0.06 | $0.003 | $0.24 |
| DeepSeek V4.1 Flash (fast)deepseek/deepseek-v4.1-flash:fast | $0.28 | $0.028 | $1.10 |
| GLM 5.3 Flashz-ai/glm-5.3-flash | $0.12 | $0.024 | $0.40 |
Prices are USD per 1M tokens. Cached input is a prompt prefix the model has seen before, which agents resend on every turn. You add credit (from $5) and each request is billed from it per token. Credits you buy never expire. Pinned model versions cost the same as their model.
Batch API
Batch requests cost the same per token as real-time requests. Upload a file of requests and collect the results within the 24 hour completion window. See the API reference.
Unlimited coding plans
One flat monthly price with no token cap on DeepSeek V4.1 Flash and GLM 5.3 Flash. Plans differ only in how many agents run at once. The first week costs about 1/6 of the monthly price.
| Plan | Agents at once | First week | Monthly |
|---|---|---|---|
| LiteOne agent, one task at a time. | 1 | $5 | $29 |
| PlusA main agent plus subagents. | 2 + extra* | $13 | $79 |
| ProA full coding setup, all day. | 4 + extra* | $27 | $159 |
| CrewSeveral projects in parallel. | 6 + extra* | $40 | $239 |
| FleetA team on one account. | 10 + extra* | $67 | $399 |
* Extra agent slots run more agents at lower priority when there is room. They are best effort and not guaranteed. Fast mode and other models are pay as you go. See the plans.
Free credit to start
New accounts get free credit to try the API. Your first top-up of $10 or more within 7 days of signing up adds $5 in credit on top.
Create an account to get an API key.
Questions
How do I buy AI tokens on OpenScale?
Create an account, add credit (from $5) and create an API key. Each request is billed per token from your credit at the prices on this page. There is no subscription and no minimum spend.
How much do 1M tokens cost?
DeepSeek V4.1 Flash costs $0.06 per 1M input tokens, $0.003 per 1M cached input tokens and $0.24 per 1M output tokens. Every model's prices are in the table above.
What is cached input?
Input tokens the model has already seen at the start of an earlier prompt, such as the system prompt and history a coding agent resends on every turn. They cost much less than new input tokens. Your usage shows how many tokens were billed at the cached rate.
Is there free credit?
Yes. New accounts get free credit to try the API. Your first top-up of $10 or more within 7 days of signing up adds $5 more.
Do credits expire?
No. Credits you buy never expire.
Is there a fee to add credit?
Card top-ups add a processing fee that covers the card network's cost, shown before you pay: $10 in credits costs $11.05 by card.
Pay as you go or an Unlimited plan?
Pay as you go bills each token and suits apps and light use. An Unlimited plan is one flat monthly price with no token cap on DeepSeek V4.1 Flash and GLM 5.3 Flash, from $29 a month, and suits coding agents that run all day. Plans differ only in how many agents run at once.
How much does the Batch API cost?
The same per-token prices as real-time requests. Upload a file of requests and collect the results within the 24 hour completion window.