Why is my model slow?
Most results arrive within 20 seconds; long outputs and multi-step tools take longer. Every plan uses the same models, so upgrading does not make a run faster.
Last updated: 2026-10-16
What's normal
Most results arrive within 20 seconds. Long outputs (a 2,000-word article, a full plan) can take a minute or more, and the text streams in as it is written. Measured over the 90 days to October 2026: half of all successful runs finished within 3 seconds, 9 in 10 within 16 seconds.
Why some runs are slower
- Length: twice the words takes roughly twice the time.
- Multi-step tools: some tools call the AI several times (an outline, then a draft); agents run several tools one after another.
- Image tools: image generation adds its own wait.
- Busy providers: Conductor picks the model per task; when that provider is slow, the run is slow.
What helps
- Ask for a shorter output if you don't need the long one.
- Keep the page open: the result streams in, and a closed tab stops the run.
- If runs are suddenly much slower than usual, check the status page.
Every plan uses the same models and the same routing, so upgrading does not make an individual run faster.
Was this helpful?
Still stuck? Contact support
