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 (849 KB).

OKInk brush dragonmax effortNeutral harness

Wall time 15m 39sCost $3.36 estimated

Review

No verdict yet.

The full review

OK. Finished on its own after 15.7 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, 3,177 characters. Every result for this prompt

[Task]
Create a 20-second motion piece at 1920x1080, 30 fps, in Chinese ink-wash style, where a dragon is painted stroke by stroke as it coils through clouds and ends on a calligraphy title and red seal.

[Scenes]
1. 0-3 s: Blank warm rice paper with visible fibres. A single ink drop falls from the top of the frame, strikes the paper and blooms outward with feathered edges, and the bloom dissolves into the first misty wash.
2. 3-7 s: Clouds are brushed in from left to right as layered spiral strokes in pale grey wash. Distant mountain silhouettes bleed in at the lowest opacity, and the camera drifts slowly upward.
3. 7-12 s: The dragon's body is painted as one long serpentine brush stroke that coils through and behind the clouds, with dry-brush scales flicking along its back. The camera tracks the brush tip along the curve.
4. 12-16 s: The head forms with whiskers, horns and mane in quick tapered flicks. In a pause, a final dot paints the eye. The dragon then surges diagonally upward, and clouds part in its wake.
5. 16-20 s: The camera settles on a vertical composition. The characters 雲龍 are written stroke by stroke down the right side, a small English line reads CLOUD DRAGON, and a red square seal presses in with a slight bounce and an uneven ink print.

[Style]
Palette: #F2EAD8 rice paper, #111111 sumi ink, #6B6B6B wash, #B9B4A6 mist, #B22222 seal red. Use a brush-script CJK font for the title and a refined thin serif for the English line. The texture is paper grain and fibre with ink pooling at stroke ends. Motion has a calligraphic rhythm: fast attack, slow taper, eased pauses between strokes and gentle spring on the stamp.

[Wow]
- Strokes have variable width driven by simulated brush pressure, with dry-brush streaks where speed is high.
- Ink bleeds outward over time along stroke edges, using a feathered, noise-driven mask.
- At least three parallax cloud layers move at different rates as the camera tracks the dragon.
- The eye-dot moment triggers a subtle ink shockwave ripple through the surrounding wash.
- The seal has carved-looking characters and an imperfect print with speckled gaps.

[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.
- The dragon and all calligraphy are built from layered, tapered brush-stroke paths, never uniform-width outlines.

[Harness notes] Render the final video to out.mp4 (1920x1080, 20 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 15m 39sFirst artifact at 14m 00s1 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 14m 03s thumbnailv1 14:03

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

Show all 149 steps

Loading the timeline…

Download the full transcript Codex CLI rollout JSONL, 420 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 $3.36, estimated from the token counts at list price.

Tokens by typeTokens$ per 1MCost
Input (uncached)219$2.00$0.0004
Cache write213,043$2.00$0.4261
Cache read10,548,227$0.20$2.1096
Output82,119$10.00$0.8212
of which reasoning48,458billed as output
Total computed from tokens
$3.3574
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 (73 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 'ink_brush_dragon' (shared by the Remotion and HyperFrames tests); all frames code-drawn; standard hyperframes harness suffix (20 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