One for All
Student · 0.5B 6 Teachers · 1.5–4B
Multi-teacher distillation

Six minds,
one student.

Six specialist AIs pour what they know into a single half-billion-parameter model — the strength of many, carried by one. Ask it anything and watch the inheritance happen, live.

watch the transfer
Live probe

One prompt. Two answers.
Before the teachers, and after.

The exact same 0.5-billion-parameter model — with and without what it learned from the six teachers. Type a question, or pick one below.

to run
Or pick a question
    After · taught by 6 teachers
    Ask the student to see how its answer changes after distillation.
    Teacher influence

    While the student answers, these bars show how much it leans on what each teacher taught it.

    idle — waiting for a prompt
    Representation space

    Where every mind keeps its thoughts.

    Each dot is how one model internally represents a piece of text. Teachers cluster by what they're good at — and the student drifts toward all of them at once. Your prompts land here as bright probes.

    Thought map

    Move your cursor — the field has depth.