oneshotlm
Compare all models on this prompt →

Matrix digital rain · qwen3.8-max

Create the Matrix digital rain effect on a full-screen black canvas: columns of falling green katakana glyphs with a trailing fade and a brighter leading character.

⛨ sandboxed Open in new tab ↗
transcript · 4 messages
user

Create the Matrix digital rain effect on a full-screen black canvas: columns of falling green katakana glyphs with a trailing fade and a brighter leading character.

assistant
thinking
Standard matrix rain canvas implementation.
write /tmp/ws/index.html (3249 bytes)
toolResult

Successfully wrote 3249 bytes to /tmp/ws/index.html

assistant
thinking
Note: the brighter head is drawn below the head's glyph — that's fine, it's one row ahead (y + FONT_SIZE). Okay. Done.

Created `/tmp/ws/index.html` — a fully self-contained Matrix digital rain effect: **How it works:** - **Full-screen black canvas** — fixed-positioned, resized on window resize, no scrollbars - **Glyph pool** — all 96 katakana characters (U+30A0–U+30FF) plus digits and a few latin symbols, picked randomly per cell - **Trailing fade** — each frame paints a semi-transparent black rect (`rgba(0,0,0,0.06)`) over the previous frame, so old glyphs dissolve gradually into long green trails - **Bright leading character** — the head glyph is drawn in near-white (`#d8ffd8`) one row below the green (`#00ff41`) body, giving each stream a glowing tip - **Organic motion** — each column has its own random speed, staggered start positions, randomized reset delays off-screen, and a 2% chance per frame of mutating its glyph so trails shimmer No external resources, no network requests — just open the file in a browser.

TAKE IT TO CODE

A single self-contained HTML file — copy it and open in any browser, or reuse the prompt with your own pi-agent run.

RUN

modelqwen3.8-max
vendorQwen
harnesspi-agent
generated2026-08-03

EVALUATION

3/5 partial

Matrix digital rain animation runs continuously but shows no distinct reaction to clicks, drags, scroll, or keyboard input.

Vision grade · anthropic/claude-sonnet-5

MEASUREMENTS

Raw readings from the capture, not a verdict — the grade above is the only judgement. Percentages are how much of the screen changed; the measure under-reports sparse motion on a dark background.

idle motion28.1%
on click30.4%
on drag30.4%
on wheel31.9%
on enter + space31.4%
on arrow keys31.4%
on w/a/s/d27.7%
frame spread14.9 / 255
console errors0
js errors none

METRICS

runtime26.7s
tokens in4.3k
tokens out1.3k
cached2.0k
cost / run$0.0130