FAQ
Common questions before you start with Drise.
Is drise.ai really using the same GLM 5.2 model as the expensive plans?
Yes. Same GLM 5.2 engine as the GLM coding plans elsewhere. We focus on one model family so we can run it cheap and run it hard. Same output quality. Bigger usage limits. A lot less money per month.
Which model variants does drise.ai expose?
Five FP8-quantised variants:
drise-glm-5.2- 1M tokens context, with reasoningdrise-glm-5.2-fast- 1M tokens context, no reasoning (lowest latency)drise-glm-5.2-short- 200K tokens context, with reasoningdrise-glm-5.2-short-fast- 200K tokens context, no reasoning (fastest)drise-vision- GLM 5.2 with vision capabilities
See Models for the full breakdown.
Are the models FP8-quantised?
Yes. Every variant is FP8-quantised for cheaper, faster inference with no meaningful loss of output quality. You do not need to configure anything; FP8 is on by default for every plan.
How is drise.ai so much cheaper than other GLM plans?
We strip the markup that other AI coding providers pass on to you. Flat monthly plans. No per-token games. We push pricing down and usage limits up. Same model. Bigger allowance. Smaller bill.
Is the API OpenAI compatible?
Yes. Drise exposes a fully OpenAI-compatible endpoint. Drop your Drise API key into any client that supports OpenAI-style endpoints (Cursor, Aider, Continue, Cline, OpenAI SDKs) and start coding. No code changes needed.
Do I get more usage than other GLM plans?
Yes. Drise plans go up to 5,000 prompts every 5 hours, starting at $0. Comparable plans elsewhere give you far less usage for much more money. Same model. Bigger limits. Lower bills.
Where can I see my usage?
Live at stats.drise.ai.