Performance

The Performance page — titled Speed & Intelligence in the app — controls how fast and how smart Zinley runs. Open Configure → Performance.

Every plan gets full quality at standard speed. Every other model is used adaptively, as a team — Zinley picks the right one for each task (see the latest lineup opens the current models). The options below make busy-hour responses faster and reasoning deeper; each is a premium upgrade billed above your standard rate.

The Performance page: Priority processing, Muse 4.1, and the Boss model chooser with Standard, Claude, and OpenAI options.
Performance — optional premium upgrades for faster, deeper responses.

Priority processing (premium)

Routes supported requests through a faster priority lane, so replies feel snappier — especially on complex, multi-step tasks. Billed at a higher rate than standard.

  • Turn on priority — the toggle on the right; switch on to speed up supported requests.
  • When you're charged — you're only charged the premium when priority is actually delivered — never for standard responses.

Muse 4.1 (premium)

An experimental Meta model that handles deep work. Falls back to GPT-5.6 Terra automatically on errors.

  • Turn on Muse 4.1 — the toggle on the right; switch on to run deep work on Muse 4.1.
  • Always completes — if Muse 4.1 hits an error, your task falls back to GPT-5.6 Terra automatically, so your work always finishes.

Boss model (premium)

When Zinley works on your tasks, one "boss" model plans the work and directs the rest of the team. Here you choose how smart that boss is — keep the standard one (included in your plan) or upgrade to a deepest-reasoning model for harder work. Upgraded bosses cost more per request. Pick one:

  • Standard — recommended (Opus 4.8 · included) — the boss every paid plan runs today: strong, fast, and proven, at your plan's standard rate. Selected by default.
  • Deepest reasoning — Claude (Fable 5 · premium) — Anthropic's most capable model; best for hard, multi-step work where planning quality matters most. Costs more per request.
  • Deepest reasoning — OpenAI (GPT-5.6 Sol · premium) — OpenAI's most capable model; a different perspective on the same hard problems, with a huge memory for long documents. Costs more per request.

The Boss model only changes who plans and directs the work — the rest of the team of models still runs as usual. Scroll for any additional deepest-reasoning options your plan offers.

How this relates to Image & video

Performance is about how Zinley thinks and replies; Image & video is about which model powers each type of media you generate.

Related