Dropstone Support

Which AI models does Dropstone use?

Dropstone runs three tiers. On the 1.8 lineup, Fast runs DeepSeek V4.1 Flash, Pro runs GLM-5.3 Flash, and Heavy runs Kimi K3, re-evaluated every month.

Dropstone runs on three models, named Fast, Pro, and Heavy. Each name is a slot in the lineup rather than a specific model: every month the best open-weight model for that job is picked, at a cost that keeps the tier affordable, and that is what the slot runs.

Available onFreeProMaxEnterpriseHeavy needs a paid plan.

So the model behind Dropstone Pro today may not be the model behind Dropstone Pro next month. The name you pick stays the same, and what sits under it improves.


What are the three tiers for?

The lineup in service is 1.8:

TierModel on 1.8Made byBest for
Dropstone FastDeepSeek V4.1 FlashDeepSeekEveryday writing, analysis, coding, quick questions. Included on every plan
Dropstone ProGLM-5.3 FlashZ.aiDifficult code, long documents, work where a first pass is not enough
Dropstone HeavyKimi K3Moonshot AILarge refactors, research across many sources, long multi-step work

Fast uses the least of your allowance, Pro a little more, and Heavy the most. All three read images and documents, and every tier draws on the same weekly allowance. How usage and limits work explains how that is measured.


Why three different makers?

The three tiers run models from three different makers, and that is deliberate. No single maker is best at every kind of work, so the lineup is assembled tier by tier rather than taken from one catalogue.

The lineup does change. When it does, the version number moves and the release notes say what changed. Dropstone lineup history is the table of every lineup so far, and Which model am I using? shows you where to look in the product.


Do I have to choose between model makers?

No, and this is the part that matters most if you have used these models directly before.

Each model above has its own platform, its own account, its own API key, and its own bill. Following them yourself means three of each, plus the work of noticing when a newer model ships, testing it against what you already use, and moving your setup across.

Dropstone does that work every month and hands you the result under one account. You pick Fast, Pro, or Heavy. There is no provider to configure, no key to paste, and nothing to migrate when the lineup moves. Your allowance and your memory are the same whichever tier you choose.


Which one should I use?

Start with Fast. It handles the large majority of tasks well and stretches your allowance furthest.

Move to Pro when Fast's answer is not good enough, or when the problem is genuinely hard.

Reach for Heavy when the task is long rather than hard: work that needs many steps, many files, or many sources.

Choose a model goes through this in more detail.


Does the model affect my memory?

No. Your memory belongs to your account, not to a model. Switch tiers mid-project and Dropstone still knows your rules, your facts, and what you have been working on.


Why these models and not others?

Dropstone runs open-weight models on purpose, and there is a plain reason for it. Why Dropstone runs open-weight models covers what open weights give you. How Dropstone chooses its models covers the monthly cycle that picks them.

Related articles

Ctrl+I