AI Model Reviewer

Claude Opus 5.5 in Claude Code 2.1.280

skills off

← Open in the gallerySame prompt, all models

Zoom, inline player and more ways to view
Poster frame of Claude Opus 5.5's page
Open the original page

The model's own page, run in a sandboxed frame only when you ask (a page that needs its own origin opens in a new tab on a separate demo site). Phones get a full-screen player capped at 1.5x pixel density (2x on larger screens) so it stays smooth; the original page runs uncapped. Tap inside for sound.

OKRetro desktophigh effortNeutral harness

Wall time 11m 52sCost $2.37 estimated

Review

No verdict yet.

The full review

OK. Finished on its own after 11.9 min.

Scores

AI judge

Not scored

No rubric for this prompt

Visitor votes and reviewer grades are not available yet.

AI judge
AI judge: the test runner (model claude-opus-5-5), scoring by visual review of the 1600 px render against the pre-registered rubric. No AI-judge score exists for this result: there is no rubric for this prompt, so no score is shown and none is made up.
Visitors
Public voting with a blind reveal is coming: you score the output first, then see the other scores.
Reviewer
The reviewer's own grade appears here once it is added.

Scores compare results for the same prompt only. AI-judge boards rank runs within one prompt, equal totals sharing a rank; scores from different prompts are never averaged.

Downloads

Files actually published for this run. Original output, site-made previews and post-run conversions are labelled separately.

The exact prompt

as run, 1,815 characters. Every result for this prompt

Build a working 1990s-style desktop operating system inside one web page. It needs a desktop with a wallpaper and icons, a taskbar with a start button, a start menu and a live clock, and real windows: each window can be dragged by its title bar, resized from its corner, minimised to the taskbar, maximised, restored and closed, and clicking a window brings it to the front. Include three working apps that open in those windows: a Notepad that edits text and saves to and reopens from localStorage under a file name; a Paint app with a pencil, brush sizes, a 16-colour palette, a flood fill and an eraser on a canvas; and a Calculator with the four operations, decimals, clear and a keyboard input path. Double-clicking a desktop icon or choosing an app from the start menu opens it. Several windows of the same app can be open at once. Everything is drawn with HTML and CSS (no images): bevelled grey buttons, a classic title-bar gradient and pixel-crisp borders. So the first view is not an empty desktop, on load the system shows a short boot screen (about 2 seconds) and then opens Notepad with a welcome note and the Calculator, cascaded. On a phone, windows open maximised and the taskbar stays usable.

Save it as index.html in the current directory: one self-contained file with inline CSS and JavaScript only, no libraries or frameworks, and no external requests of any kind (no CDNs, web fonts, images or audio files). It must work when opened directly from disk, fit any window from a 390 px wide phone to a desktop, and run without console errors. You may serve it locally and open it in headless chromium (/usr/bin/chromium, software WebGL, no GPU) to screenshot and inspect your own output, and iterate on your own before finishing. Close any browser you open as soon as you've taken your screenshot.

How it was made

Wall time 11m 52sFirst artifact at 8m 10s3 saved versions

Versions saved during the run

3 intermediate outputs, in the order the model wrote them.

Version 3 of 3
  1. Version 1 at 8m 13s thumbnailv1 8:13
  2. Version 2 at 8m 23s thumbnailv2 8:23
  3. Version 3 at 10m 24s thumbnailv3 10:24

Step by step

The prompt, every agent turn and tool call (tool name and a one-line summary; inputs and outputs are not shown), then the final message. The full machine-format transcript is published as a scrubbed download.

Show all 34 steps

Loading the timeline…

Download the full transcript Claude Code session JSONL, 464 KB. Scrubbed: anything that looks like a key, token, e-mail address or phone number is replaced, and account details and the test machine's file paths are removed.

Cost breakdown

Headline cost $2.37, estimated from the token counts at list price.

Tokens by typeTokens$ per 1MCost
Input (uncached)36$4.00$0.0001
Cache write118,167$5.00$0.5908
Cache read1,637,142$0.20$0.3274
Output72,784$20.00$1.4557
of which reasoning25,937billed as output
Total computed from tokens
$2.3741
Total reported by the harness
none

Computed from harness-reported token counts x list prices.

Prices: Project price table (vendor list prices), 2026-09-22.

Run notes

one prompt, agentic harness (self-review allowed)labelled re-run (of appclone_retro_desktop__claude-code__opus-5-5-high)

  • frozen prompt (written): written 2026-10-01 for the MR new-categories expansion (owner request: expand model review with new kinds of tests); category app_clones; zero results
  • plain wave-1-equivalent cc_config copy (settings.json + .claude.json; no ECC files)
  • Mode: interactive TUI in tmux (160x48), inside bubblewrap sandbox
  • Context setting: 1M ([1m] model name)
  • Isolation audit: none