AI Model Reviewer

GPT-6 Astra in OpenCode 1.18.32

skills off

← Open in the gallerySame prompt, all models

Zoom, inline player and more ways to view
Poster frame of GPT-6 Astra'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.

OKCode editormax effortNeutral harness

Wall time 21m 41sCost $5.48 reported

Review

No verdict yet.

The full review

OK. Finished on its own after 21.7 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, 2,012 characters. Every result for this prompt

Build a working code editor in one page in the spirit of a lightweight VS Code. It needs: a file explorer sidebar with a small seeded project (index.html, style.css, app.js, utils.js, README.md) where files can be created, renamed and deleted; tabs for open files with unsaved-change dots; an editor component you write yourself (no libraries) with line numbers, syntax highlighting for JavaScript, CSS, HTML and Markdown, auto-indent, bracket matching and auto-closing, multi-line indent and outdent of selected lines from the keyboard, current-line highlight, and a minimap; find and replace with match count and regex toggle; a command palette (Ctrl+Shift+P) with fuzzy search over commands and files; a run button that executes app.js in a sandboxed iframe with the project's HTML and CSS and shows its output plus console logs and errors in a bottom panel; light and dark themes; and persistence of the whole project to localStorage. So the page is alive without input, on load it runs a scripted demo: the cursor opens app.js, types a small function character by character with realistic pauses, autocloses brackets, uses find and replace to rename a variable, opens the command palette, runs the project so a little animation appears in the preview and a log line in the console, then reverts and repeats with a different snippet, until the user touches anything. Make it feel fast and professional.

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 21m 41sFirst artifact at 11m 02s5 saved versions

Versions saved during the run

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

Version 5 of 5
  1. Version 1 at 11m 10s thumbnailv1 11:10
  2. Version 2 at 17m 04s thumbnailv2 17:04
  3. Version 3 at 17m 39s thumbnailv3 17:39
  4. Version 4 at 19m 51s thumbnailv4 19:51
  5. Version 5 at 21m 17s thumbnailv5 21:17

No transcript was delivered with this result.

Cost breakdown

Headline cost $5.48, as reported by OpenCode (its own bill for the run). The list-price estimate from the token counts is $5.31.

Tokens by typeTokens$ per 1MCost
Input (uncached)123$10.00$0.0012
Cache write70,147$10.00$0.7015
Cache read1,966,736$1.00$1.9667
Output52,778$50.00$2.6389
of which reasoning12,010billed as output
Total computed from tokens
$5.3083
Total reported by the harness
$5.4837
Difference
Harness-reported $5.4837 vs list-price estimate $5.3083 (-3.2%); the harness figure is the headline.

OpenCode 1.18.32 reported its own total; token counts x list prices shown beside it. Cache writes priced at the input rate (the price table has no separate cache-write price). The harness reports reasoning tokens separately from output; here output includes them (40,768 visible + 12,010 reasoning), and both bill at the output rate.

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

Run notes

one prompt, agentic harness (self-review allowed)

  • frozen prompt (written): written 2026-10-05 for the MR category expansion (owner request: 500 finished runs in every category); category app_clones; zero results
  • plain OpenCode config copy (providers only; no ECC instructions/agents)
  • opencode export written to a file (not piped)
  • Mode: interactive TUI in tmux (160x48), inside bubblewrap sandbox
  • Isolation audit: outside paths seen
  • Renderer: Canvas 2D (getContext('2d') at index.html line 330) plus DOM/CSS UI and inline SVG; single self-contained index.html, no CDN and no external assets. The published copy renames one CSS class (the words tab and bar written as one word) to "tab-bar" (7 occurrences, CSS and markup only, no script use) to pass the site's banned-string gate; nothing else is changed.
  • Card video: the run's own session capture (headless Chromium, no GPU), re-encoded as H.264. Poster: the run's own render.png.
  • Same frozen prompt as the other rows on this test; configurations differ and all rows are unscored, so this is not a head-to-head result.