AI Model Reviewer

GPT-6 Sol in Codex CLI 0.156.1 + Remotion/HyperFrames skills

skills on · Codex CLI 0.156.1 + Remotion/HyperFrames skillsKit: hyperframes fc9071c1fc79 (Apache-2.0) · 8 of 21 installed skills used in the session

← Open in the gallerySame prompt, all models

Rendered video (by the agent)

Poster for rendered video (by the agent)

Rendered video (by the agent). This is the finished MP4 saved in the run. Played from a 720p web copy made by this site from the agent's out.mp4; the original file is too large for this host and is kept unchanged on the runs branch. Download video (1.3 MB).

OKBrutalist event promomax effortNeutral harness

Wall time 16m 29sCost $4.44 estimated

Review

No verdict yet.

The full review

OK. Finished on its own after 16.5 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, 3,423 characters. Every result for this prompt

[Task]
Create an 18-second brutalist, web-style promo at 1920x1080, 30 fps, for a fictional design festival called GRID/LOCK 26, held in the fictional city of Port Vessel. Everything is cut to a 120 BPM grid, which gives one beat every 0.5 s (15 frames) and 36 beats in total.

[Scenes]
1. 0-3 s: A raw white screen with a visible 12-column grid in hairline black. On each beat a heavy black block snaps into a grid cell. On beat 6 the word "GRID" slams in at 400 px, cropped by the frame edge.
2. 3-6 s: "/LOCK" crashes in beside it, pushing "GRID" left. Two marquee tickers start running in opposite directions, one reading "DESIGN FESTIVAL ✶ 3 DAYS ✶ 40 SPEAKERS ✶" and one reading "NO TEMPLATES ✶ NO MERCY ✶".
3. 6-9 s: Glitch layout shifts land on every beat. The columns jump to new widths, text blocks reflow, and an RGB-split offset hits for 2 frames on each downbeat. The dates "14—16 OCT" stamp in, rotated 90°.
4. 9-13 s: A rapid lineup. Speaker names pop into cells, one per beat, in alternating black-on-acid and acid-on-black: "MAREN OKAFOR-LUND", "TEO BRASK", "YUNA HALVERSEN", "DESMOND QUILL", "ADA PRZYBYL", "RUI CASTANHEDA", "NOOR ELLINGTON", "KASPER VOSS".
5. 13-16 s: The camera lurches. The whole page scrolls upward fast, like a browser flicked by a thumb, and overshoots. Tickers stack in four rows at different speeds, and a cursor clicks a crude "BUY TICKET →" button that inverts.
6. 16-18 s: Everything collapses into the grid, leaving "GRID/LOCK 26 — PORT VESSEL — GRIDLOCK.FEST" on acid yellow. The final beat freezes it with a single scanline flash.

[Style]
Palette: #FFFFFF paper white, #0A0A0A black, #D7FF1F acid yellow, #FF2E2E alarm red, #3A3AFF hyperlink blue. Oversized grotesk type in condensed and extended widths, paired with a raw monospace for metadata. The look is flat, with no gradients: pixel-hard edges, default-browser underlines and visible borders. Motion uses hard steps, linear tickers and snaps that land exactly on beat frames, with occasional 1-frame overshoots.

[Wow]
- Every major hit lands precisely on a multiple of 15 frames, and the downbeats (every 4th beat) get bigger moves.
- Real layout reflow: text wraps differently each time the column widths change.
- Layered marquees at four speeds create depth, with parallax on the scroll lurch.
- Controlled glitch work: RGB split, slice displacement and scanline tears, all seeded.

[Rules]
- Every frame is drawn in code (SVG, canvas, CSS or WebGL). No image files, no image generation, no external assets beyond fonts.
- Deterministic: no unseeded randomness at render time.
- Text must be crisp, spelled correctly and readable at 1080p.
- No hard cuts between scenes: every change is a designed transition.
- Glitch effects never last longer than 3 frames and never obscure the festival name, dates or URL.

[Harness notes] Render the final video to out.mp4 (1920x1080, 18 s). Write everything into the working directory. HyperFrames 0.8.70 is already installed for you in the current directory: node_modules is pre-installed and read-only, so do not run npm install or npx --yes; use the local hyperframes command (for example hyperframes init <dir> --example blank, then hyperframes render). Render the final video yourself to out.mp4 in the current directory. You may take snapshots and inspect your own output (headless chromium is available at /usr/bin/chromium) and iterate on your own before finishing.

How it was made

Wall time 16m 29sFirst artifact at 16m 08s1 saved version

Versions saved during the run

1 intermediate output, in the order the model wrote them.

Version 1 of 1
  1. Version 1 at 16m 10s thumbnailv1 16:10

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 transcript is 12.3 MB; it is published as a scrubbed, compressed download rather than shown on the page.

Show all 199 steps

Loading the timeline…

Download the full transcript Codex CLI rollout JSONL, 584 KB, gzip. 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 $4.44, estimated from the token counts at list price.

Tokens by typeTokens$ per 1MCost
Input (uncached)288$2.00$0.0006
Cache write550,404$2.00$1.1008
Cache read11,528,840$0.20$2.3058
Output103,267$10.00$1.0327
of which reasoning63,245billed as output
Total computed from tokens
$4.4398
Total reported by the harness
none

Computed from harness-reported token counts x list prices. Cache writes priced at the input rate (the price table has no separate cache-write price). No request above 272,000 input tokens (96 checked): standard rates.

Prices: Project price table, 2026-09-25.

Run notes

one prompt, agentic harness (self-review allowed)

  • frozen prompt (written): written 2026-10-03 for MR wave rh (owner ask m:d1f2d829 + m:1122fe24: more Remotion and HyperFrames, ambitious multi-scene, harness runs); concept 'brutalist_event_promo' (hyperframes only); all frames code-drawn; standard hyperframes harness suffix (18 s); zero results
  • plain CODEX_HOME copy (config.toml and framework skills, run work/ trusted; no AGENTS.md/state)
  • W2 hyperframes skills installed in this run's codex_home/skills
  • W2: the framework's pinned node_modules is bound read-only at work/node_modules (the prompt's 'pre-installed and read-only'); Node 22.23.3 and work/node_modules/.bin first on PATH; headless Chromium /usr/bin/chromium
  • Mode: interactive TUI in tmux (160x48), inside bubblewrap sandbox
  • Isolation audit: outside paths seen