oneshotlm
Compare all models on this prompt →

Reaction-diffusion · nemotron-3-ultra-550b-a55b

Output generated Cost / run $0.0298 Runtime 60.4s

Prompt for this run

Run a Gray-Scott reaction-diffusion simulation on a canvas that grows organic Turing patterns in real time, with the mouse seeding new growth.

⛨ sandboxed Open in new tab ↗

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

modelnemotron-3-ultra-550b-a55b
vendorNVIDIA
harnesspi-agent
generated2026-07-22

METRICS

runtime60.4s
tokens in24k
tokens out9.7k
cached8.3k
cost / run$0.0298
Transcript · 6 messages
user

Run a Gray-Scott reaction-diffusion simulation on a canvas that grows organic Turing patterns in real time, with the mouse seeding new growth.

assistant
thinking
The user wants a Gray-Scott reaction-diffusion simulation as a single-page web app in index.html. This is a classic Turing pattern simulation where two chemicals U and V react and diffuse, creating organic patterns. The mouse should seed new growth.

Let me create a complete, self-contained HTML file with:
1. Canvas for rendering
2. Gray-Scott simulation using GPU (WebGL) or CPU (JavaScript) - I'll use WebGL for performance
3. Mouse interaction to seed patterns
4. Real-time animation

I'll write this as a single HTML file with inline CSS and JS.
→ write /workspace/index.html (25583 bytes)
toolResult

Successfully wrote 25583 bytes to /workspace/index.html

assistant
thinking
The file has been created. Let me verify it works by checking the file exists and its content.
→ bash {"command":"ls -la /workspace/index.html"}
toolResult

-rw-r--r-- 1 node node 25593 Jul 22 17:46 /workspace/index.html

assistant
thinking
The file is created. The web app is complete with:
1. Gray-Scott reaction-diffusion simulation using WebGL 2 for GPU acceleration
2. Real-time Turing pattern generation
3. Mouse/touch interaction to seed new growth
4. Adjustable parameters (Feed, Kill, Diffusion rates, Speed)
5. 9 classic presets (Mitosis, Coral, Worms, Maze, Spots, Chaos, Fingers, Spirals, Replicators)
6. 7 color palettes (Classic, Ocean, Fire, Neon, Mono, Viridis, Plasma)
7. Reset and palette cycling buttons
8. Responsive design that works on mobile
9. Single self-contained HTML file with inline CSS and JS

The implementation uses:
- Ping-pong framebuffers for simulation
- 9-tap Laplacian for isotropic diffusion
- 1D texture for color palette mapping
- Full-screen quad rendering

This should run smoothly in any modern browser with WebGL 2 support.

Created `/workspace/index.html` — a complete Gray-Scott reaction-diffusion simulation with: - **GPU-accelerated WebGL 2** simulation (512×512 grid, ping-pong framebuffers) - **Real-time Turing patterns** — spots, stripes, worms, chaos, spirals, replicators - **Mouse/touch seeding** — drag to inject V-chemical and grow new structures - **9 classic presets** (Mitosis, Coral, Worms, Maze, Spots, Chaos, Fingers, Spirals, Replicators) - **7 color palettes** (Classic, Ocean, Fire, Neon, Mono, Viridis, Plasma) with cycling button - **Live parameter control** — Feed (F), Kill (k), diffusion rates, simulation speed - **Reset button** to re-seed from center - **Responsive** — works on desktop and mobile Open `index.html` in a browser to run.