Jelly physics simulation improvement comparison
- Views
- 79K
- Likes
- 360
- Replies
- 22
- Saves
- 162
An interactive WebGPU 'Citrus Jelly' scene, a soft-body orange slice you can grab and stretch, built by Opus 5.5 from a detailed prompt on appearance, physics and UI; costs compared across versions.
Full prompt2,490 chars
Create Citrus Jelly — an interactive 3D scene featuring a gummy orange slice. Everything must run directly in the browser in a single HTML file using WebGPU and WGSL shaders. Build genuine volumetric 3D geometry, not an image or a prerecorded animation. APPEARANCE A thick semicircular orange slice with vibrant orange flesh, a pale pith layer, and a firmer outer rind. Visible radial segments, thin internal membranes, and tiny bubbles. A juicy, translucent gummy material with soft highlights, light refraction, and depth-dependent color. Avoid making it look like rigid plastic or glass. Use a light cream background, soft studio lighting, and a contact shadow. Avoid overexposure. PHYSICS AND INTERACTION Let users grab any visible point with a mouse or finger and pull it. Deformation must originate at the grab point and propagate naturally through neighboring areas. The flesh should stretch and narrow under tension, while the rind offers more resistance. On release, the slice should spring back gently, with oscillations gradually settling. Use a stable soft-body simulation: a spring mesh or volumetric elastic constraints with volume preservation. Segments, membranes, and bubbles must deform with the flesh rather than separating or floating independently. Include ground collisions: the slice should squash slightly on impact and wobble afterward. Prevent instant snapping back, endless oscillation, and exploding geometry. INTERFACE A minimal editorial layout with generous whitespace. A large italic serif heading, “Citrus Jelly.”, on the left. A compact control panel on the right with Firmness and Internal damping sliders. Give it a nudge, Reset, and Pause buttons. Show mesh and ¼ speed toggles. Three color presets: Orange, Blood Orange, and Lemon. A small interaction hint: “Grab. Stretch. Let go.” All labels in English, with comfortable desktop and mobile controls. TECHNICAL REQUIREMENTS Genuine WebGPU rendering, not a Canvas 2D substitute. Generate geometry and materials in code without premade models. Use a fixed simulation timestep and constraints that prevent excessive stretching. Maintain smooth performance during grabbing, dragging, and release. Display a clear fallback message when WebGPU is unavailable. Test different grab locations, strong stretching, oscillation decay, collisions, and reset. The main priority is a convincing gummy feel. The slice should look soft, juicy, and so satisfying in motion that people want to keep stretching it.
Source: the creator's reply under the post. Kept in its original language, unedited.
Want more prompts like this one?
Get the 50 most-viewed Claude 5.5 video prompts on one page: the full original text, the creator and the post it came from. We'll only email you when there's a new prompt worth trying. Unsubscribe anytime.
Original post: https://x.com/vib3coded/status/2103822857155313953
All works are open data at GitHub · awesome-opus-5.5-video.
Model attribution is as stated by the creator and was not independently reproduced. View counts are a snapshot from 2026-11-13. The work belongs to its creator; DM @GoSailGlobal on X for corrections or removal.
More comparisons
0:42AI self-portrait challenge: Sonnet 5.5 vs GPT-6 Astra
7.3M views
0:28Opus 5.5 vs GPT-6 Astra samurai game demo
1.1M views
0:13Gummy bear soft-body physics scene
632.1K views
0:30GPT-6 Astra vs Claude Opus 5.5 comparison
605.2K views
0:22Grok vs Opus grass-touching simulator cost test
555.4K views
0:44Opus 5.5 vs GPT-6 Sol Unreal game dev
544.7K views
Related articles
OpenAI Academy vs Anthropic: A Complete Comparison of Two Leading AI Companies' Free Learning Paths
Both of the top AI companies offer official free learning platforms — together that's 25+ courses and hundreds of hours of content, enough to take you from AI newcomer to resume-worthy engineer. This post maps out all 15 OpenAI Academy courses along a 10-week roadmap and provides a full side-by-side comparison with the Anthropic learning path.
Anthropic Claude Certified Architect Complete Roadmap: 8 Free Courses + $99 Exam
Anthropic launched the Claude Certified Architect certification on March 12, 2026, with 13 self-paced courses — all completely free, no registration required, no paywalls. The exam itself costs just $99. This post walks through a 12-week roadmap covering the 8 key courses in Weeks 1–9, with direct links to each.
OpenAI Agents SDK Major Upgrade: Harness and Sandbox Now Built Into the SDK — The Opposite Approach from Anthropic
OpenAI has shipped a major upgrade to its Agents SDK, with the headline changes being harness and sandbox support built directly into the SDK. Anthropic's Claude Agent SDK takes exactly the opposite approach — one gives you everything pre-assembled, the other hands you a toolbox.
Meta's Official PyTorch Tutorials: The De Facto 'Deep Learning Field Manual' — 9 Steps from Training to Deployment
Meta never built an AI academy, but it turned the official PyTorch tutorials into an industry standard — the de facto deep learning field manual. Completing them is the equivalent of working through a full modern deep learning curriculum. This post maps out the recommended 9-step path plus Llama companion resources.