Cclssai
CLSSAI gateway

Model Lab

Run real model requests through the authenticated CLSSAI gateway.

599 OpenRouter models
First successful run1 / 4 setup steps ready

Start with an exact model and mode, add the selected execution key, keep the sample, then run through CLSSAI gateway.

  1. 1
    ModeImage understanding
  2. 2
    Real-run modelOpenAI: GPT-4o-mini · verified 2026-08-05
  3. 3
    CLSSAI API keyPaste your CLSSAI API key below
  4. 4
    Sample inputAdd an image file to complete the sample

The key stays only in this tab's React memory, is sent through the authenticated CLSSAI Worker only for requests, and is cleared by refresh or tab close. Runs use your CLSSAI gateway account and billing.

Task presetsModel + parameters + system prompt · browser-local

Presets never store the temporary Key, user prompt, uploads, PDF parses, generated media, or results. Open applies a preset on the other execution page.

Vision model273 compatible

google/gemini-3.5-flash-lite:batchtext + image + video + file + audio → text

Browser-local · shared with model details and Chat
Attachmentrequired
ReasoningRequired by model

Reasoning tokens count as billed output tokens. The provider's explicit summary is shown separately from the final answer; raw or encrypted traces are never relabeled as a summary.

OpenRouter/provider decides · published default minimal · enabled by default.
Request tools (server-side) Beta2 active

The model chooses whether to call an enabled tool. This run is capped at 4 server-tool calls; official model and tool usage may apply.

Web Search and Web Fetch start active. Changes made after a result apply only to the next run; the result keeps its own frozen tool snapshot.

Inputs and outputs are not saved by this Playground.