AI Model Reviewer

Claude Opus 5.5 in Claude Code 2.1.288

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. It starts in its auto-tour (T switches to play mode on a keyboard).

OKSanta Monica particle beach (Opus 5.5 demo prompt)default effortNeutral harness

Wall time 22m 26sCost $4.01 reported

Review

No verdict yet.

The full review

OK. Finished on its own after 22.4 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

agentic harness (verbatim prompt + deliverable line), 493 characters. Every result for this prompt

Create a stylistic 3D environment of a busy Santa Monica beach that adapts to the time of day. The art style must be a particle illustration, which uses thousands of tiny glowing specks rather than solid fills with flowing ribbon strokes behind figures to suggest motion.

---
Deliverable: build this in the current folder as a self-contained web page that opens in a browser with index.html as the entry point (no build step; CDN imports are fine). Work autonomously and do not ask questions.

How it was made

Wall time 22m 26s

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 56 steps

Loading the timeline…

Download the full transcript Claude Code stream JSON, 432 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 $4.01, as reported by Claude Code (its own bill for the run). The list-price estimate from the token counts is $4.01.

Tokens by typeTokens$ per 1MCost
Input (uncached)66$4.00$0.0003
Cache write188,474$5.00$0.9424
Cache read4,260,706$0.20$0.8521
Output110,609$20.00$2.2122
of which reasoning51,773billed as output
Total computed from tokens
$4.0070
Total reported by the harness
$4.0070
Difference
matches

Claude Code 2.1.288 reported its own total; token counts x list prices shown beside it.

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

Run notes

one prompt, agentic harness (self-review allowed)Opus 5.5 demo prompt (awesome-opus5-5-videos, MIT)re-run after infra timeout

  • Prompt from the open-source collection awesome-opus5-5-videos (github.com/yihui-dev/awesome-opus5-5-videos, MIT License, (c) 2026 yihui-dev), originally posted by @ishuagra02: https://x.com/ishuagra02/status/2102920408743678129
  • Prompt sent verbatim, plus one fixed deliverable line (build index.html in the run folder, no build step, CDN imports allowed, work autonomously). One scored attempt per harness; this is the 2nd run of this cell after an infra timeout (see next note).
  • The original post's text is trimmed to its first paragraph; the rest was the author's note about the video, not part of the prompt.
  • Run with no time or turn limit. An earlier attempt of this cell was stopped by our own 30-minute timeout on an overloaded machine while the agent was testing its page; that attempt is not shown (it was our infrastructure, not the model).
  • Capture: 1280x720 headless Chromium, software WebGL (no GPU); capture fps 2.7 is a software-render figure, not a real-device frame rate. Console errors during capture: 0.
  • Mode: headless (claude -p, stream-json)