Skip to content
All models

Clef Omni

A decision model that reads text and images and returns a choice, a score, or a yes/no answer.

Model info

  • Provided by: system-one.dev
  • 4 credits per question per 4KB
  • Up to 32 questions · 13MB per request
  • Inputs: text, JSON, and images (PNG, JPEG, WebP)
  • Images: up to 4 · 4MB each · 8MB total
  • Pick one: up to 255 options; Score: 2–10 levels
  • Released 2026-10-09

Uses

  • Ticket routing with screenshots: Read the user’s description and screenshot together, then pick a team or a tool. See Agent tool routing.
  • Text-and-image moderation: Read a post and its image, then allow, block, or hand off from a yes/no probability. See Agent guardrails.
  • Screenshot classification and routing: Look at an uploaded screenshot and decide what kind of problem it is, or which model should handle it. See Model routing.
  • Text-and-image scoring: Use Score on a generated image or a text-and-image result, as a first-pass evaluation. See Using Score instead of an LLM as judge.
  • Agent completion checks: Look at a UI screenshot and decide whether the task is done. See When an agent should stop.

Official sources

Open GET /v1/models · Developer docs · Try live (sign in required)

Frequently asked questions

Is Clef Omni a large language model?
No. It does not generate text. You ask a few questions (pick one, score, or yes/no) and it only returns a choice, a score, or a probability. Use a generative model when you need written content.
What inputs does Clef Omni accept?
On this site it currently accepts text (including JSON) and images. Limits are in Model info on this page.
How is Clef Omni different from Clef and Clef Flash?
The three share the same request format and all accept text and images. Clef Flash is the cheapest and fits high-volume, cost-sensitive work. Clef has the highest rate; when Jev or Clef Flash is not good enough, compare Clef on the same data. Clef Omni sits between them. Compare the rates in each page’s Model info.
Does Clef Omni support Chinese?
Test Chinese results with your own data first.
How is it billed?
By credits, from the number of questions and the request size. See Model info on this page. Details: Pricing.
Can I try it live?
Yes. Try live (sign in required).