Kayts

All models

Kayts Nareia

Nareia: a name from Armenia, for the long view

Built for long work: Kayts Nareia keeps up to a million tokens in view, so whole reports, books and long conversations stay in context from the first message to the last.

  • From the Pro plan

What it can do

Text only, for now

This model reads text and files, not images yet. Kayts Mihr and Kayts Anahit read images in the same chat.

Thinking mode

For harder questions it reasons before it answers; you choose how long it may think.

Research tools

Looks up papers on its own, reads the links you send, and cites its sources.

1M tokens of context

How much it keeps in view at once: your conversation, your files and the sources you add.

Kayts Voice Mode

Talk with it and hear the answers in a natural voice, in every plan.

Benchmarks

How Kayts Nareia scores on public tests, as published by the company that makes the model.

GPQA DiamondScience questions at PhD level 90.1%
Humanity's Last ExamExpert questions, without tools 37.7%
LiveCodeBenchCoding problems 93.5%
SWE-bench VerifiedFixing real bugs in open-source projects 80.6%
BrowseCompFinding hard-to-find facts on the web 83.4%
MRCR 1MRecalling details across a million tokens 83.5%

Scores published by the company that makes the model behind Kayts Nareia, from its own tests in its deepest thinking mode; BrowseComp with tools. Which company that is, and the published source, is in the Privacy Policy. Benchmarks and their versions differ between models, so compare only the same benchmark and version. Checked 20 October 2026.

Best for

  • Very long documents and whole books
  • Projects that span many conversations
  • Research over many sources at once

Good to know

Kayts Nareia reads text and files, not images yet. It comes with the Pro and Max plans.

Our models carry Armenian names: Kayts (կայծ) means spark. Which companies run them is listed in the Privacy Policy.

The other models