One assistant.
Four depths of work.
Every LumenQube tool runs the same assistant. What you choose is how far it goes to be right — a single fast pass, a build it verifies, a build it proves a second way, or an engine that interrogates your data before it trusts it.
The same quarter close, run four ways. The job never changes — only how far the engine goes before it calls the number finished.
Depth is a shape, not a slider.
The engine does not change what the assistant is. It changes how many of the same five steps it gets, how many times it checks itself, and how much room it has in one turn.
Every run is the same five steps.
Gather, plan, create, review, enhance. Fast takes the short road; the deeper engines add the plan and then loop the last two until the result stops changing.
- Nothing touches your file until the plan exists
- Review reads the finished result back out of the document
- Enhance fixes what review found — and re-derives the headline numbers
The same four depths, in every editor you own.
Read down a column to see what changes as you move up. Every tool takes every engine your plan includes — the picker sits in each editor’s AI panel from Plus up.
- One assistant across PDF, sheets, docs, slides, design, notes and forms
- Each editor sets its own step budget; a deeper engine gets more room in all of them
- Each plan has a default engine, so you rarely have to choose at all
One allowance. One rate. No premium tier for being right.
You are never charged a different rate for a deeper engine. It simply takes more steps to get there, so it uses more of the one monthly allowance you already have.
- Every account gets a free allowance each month
- Cap what a single run may spend — at the ceiling it stops and reports
- Fast for the dozens of small asks; the deep engines for the jobs that earn it
Pick the depth, not the model.
You are never charged a different rate. A deeper engine simply uses more of the one monthly allowance per step, and gives the work more room.
Lumen Fast
The dozens of small asks in a normal day. Answers land almost immediately and use the least of your month.
Lumen Core
Real document work. It holds a long file in mind and follows multi-part instructions, so the first draft is usually the one you keep.
Lumen Expert
Work where being wrong is expensive. It thinks before answering, then proves the answer instead of asserting it.
Lumen Ultra
The hardest work you have. Our most capable engine, for jobs where deciding what to do is harder than doing it. Max can select it; Ultra runs on it by default and can switch it into its deepest reasoning.
Step budgets shown are LumenSheets’. Every editor sets its own, and a deeper engine always gets more room in all of them. You rarely have to choose at all — each plan has a default engine, and the picker sits in every editor’s AI panel from Plus up.
The engines · common questions
Do I have to pick an engine every time?
No. Each plan has a default engine and the assistant uses it automatically. The picker is there for the moments when one particular task deserves more depth than your usual default.
What does “checks its own work” actually mean?
After building something, the engine reads the finished result back out of the real document — the computed values in a sheet, the rendered page in a PDF, the board as it actually draws — and fixes whatever does not hold up. Lumen Fast skips this and delivers one confident pass. Lumen Core does it once. Lumen Expert and Lumen Ultra repeat it until the result is clean, and additionally re-derive the headline numbers a second, independent way.
Which engines does my plan include?
Free includes Lumen Fast and Lumen Core, with Fast as the default so a free month is not spent three times faster than you expected. Plus makes Core the default. Pro adds Lumen Expert on demand and Max makes Expert the default. Max also reaches Lumen Ultra whenever you pick it — deliberately not as the default, so a Max month is not spent three times faster by somebody who never opened the picker. Ultra runs on Lumen Ultra by default and is the only plan that can switch it into extended thinking. You can change plans at any time.
Does a deeper engine cost more money?
Not a different rate — usage is priced from what a request actually costs to serve, so a deeper engine simply uses more of your monthly allowance per step. Your plan price does not change.
Which engine does LumenAgent use?
The Agent uses your plan’s engine too. Deeper engines matter most here, because an autonomous run makes many decisions in a row and each one compounds.
Do my documents get used to train these models?
No. Your content is sent only when you ask for an AI action, only the content needed for that action is sent, and it is never used to train models. See Security for the details.
Every engine, in one app.
Download free and you get Lumen Fast and Lumen Core, with a monthly allowance included. Move up whenever the work asks for it.