Three.js Prompt

Claude Opus 5.5 on our AI coding benchmark: build a fully autonomous 3D water simulation from an empty repo - procedural mountains, rain, streams, lakes. TypeScript + Three.js, no physics engine. One prompt, zero corrections, ~21 min. 23/23 tests green. https://github.com/AiondaDotCom/ai-sim-benchmark

One prompt. One interactive camera lab. 🎥🔥 Claude Opus 5.5 generates a 3D experience where you can adjust the focus ring, move lens elements, shift the focal plane, and see real-time depth-of-field effects. AI-generated interactive 3D is getting wild. 🤯

I asked Opus 5.5 to explain gravitational lensing by building an interactive black hole lab One prompt for the first version, then a dynamic workflow to polish it for three rounds while I slept. This is what I woke up to: 5h 28m end to end, $90 at API prices https://blackhole.voxyz.ai Drag the black hole and the sky behind it bends into arcs, closing into a full ring when everything lines up Each round ran 3 agents. Two reviewers each picked 9 to 10 issues and ranked them P0, P1, P2. The third one fixed them. Round 1: 3D materials and lighting + UI and type Round 2: camera motion and interaction + science-communication design Round 3: a design director deciding whether it ships + visual QA and performance Every P0 got fixed, the rest only if worth it. When the two reviewers disagreed, it picked one and wrote down why. It backed up before each round, checked screenshots after, and rolled back anything it couldn't fix.

Opus 5.5 generated this simulation of molecular orbitals in about ~20 minutes. I had to prompt for some structures and resonance but I can see the promise of instant educational simulations for any advanced topic

110k rigid bodies in real time, in a browser tab. Grab anything, fire cannonballs. Built with Claude Opus 5.5: Augmented Vertex Block Descent (SIGGRAPH 2025, Giles, Diaz & Yuksel) in WebGPU with @threejs . It's a physics solver that stays stable with huge scenes and big mass ratios. Opus 5.5 even put together this whole demo reel video just from a single text prompt in Claude Code haha 🤯 Try it: https://three-avbd.vercel.app (recorded on an M4 Max, lighter scenes run fine on older computers or on mobile) Code's here, MIT licensed. Go build something with it, have fun breaking things: https://github.com/sbobyn/three-avbd #threejs #WebGPU

Asked claude Opus 5.5 to make an explainer on how a mechanical keyboard works. It wrote the music in code, used real switch recordings as the beats, built the 3D in @threejs and timed every cut to the beat. No stock footage. no editing app. What a time to be alive!

I TRIED MAKING A 3D MOTION GRAPHIC WITH CLAUDE OPUS 5.5. THIS IS WHAT CAME OUT. The idea was simple: start at my desk and don't stop zooming until we hit a single atom inside the CPU. Opus 5.5 took it and just kept going. Into the PC case. The cooler lifts off. The CPU under it. The metal lid comes off. Then the silicon die, one core, the copper wiring stacked on top, rows of logic cells, the fins and gates of the transistors… …all the way down to one silicon atom, bonded to its four neighbours. One continuous shot, no cuts. The ruler on the right tracks the scale the whole time: from about a metre down to about 2 nanometres. And the part that still gets me: every frame of this is 3D rendered right in a browser, from code. No Blender, no After Effects, no stock footage. A year ago this would have been a studio job. Now it's one conversation. Opus 5.5 is on another level.

From Cabo da Roca to observable universe. I gave opus 5.5 old prompt I prepared for Fable but I was not satisfied with the results. I asked model to create a cinematic film that zooms out from a hillside on Earth to the full observable universe using @threejs No external assets. Time: ~3h Output tokens: 662,769 Cache: 125.6M read, 2.0M written; API-equivalent cost: about $54.52: output $13.26, cache reads $25.12, cache writes $16.14

I asked opus 5.5 in @agnt_gg for the most realistic fire it could build using ONLY javascript. It wrote a real 3D fluid simulation that runs on the GPU. The flame colors come from blackbody physics, the smoke is real volume, it runs at 96 fps in 1080p, and it critiqued its own renders about 40 times. Then it shot a 38 second film of it. The sound comes from the fire itself. 🔥 Details: - Nothing faked. The fire is actual simulated gas that burns, heats up, rises, and swirls. It gives off soot that glows, and the soot that escapes cools into smoke. Every frame is computed on the GPU. - The AI graded its own work: render a frame, have a vision model pick it apart like a VFX supervisor, fix it, repeat. It caught a flame that kept dying, ring shaped artifacts on the ground, and coals that looked like cheese chips. - The film isn't screen recorded. Each frame is rendered offline with scripted camera moves. The crackle and roar come from how brightly the fire burns in each frame. - No Blender, no stock footage, no sound library. One HTML file, one conversation.

I asked Claude Opus 5.5 to explain how a rocket engine works by building an interactive Raptor 3 you can take apart in your browser. Cut it open, follow the oxygen and the methane through both turbopumps, then throttle it and watch the shock diamonds move. http://airsup.ai/rocket-engine

Opus 5.5 just turned a smartphone into a 20-second trip from glass to atoms. It starts on the phone, dives through the motherboard and chip, then keeps going through the semiconductor structure until you’re basically staring at the atomic scale. This is the kind of 3D shot that normally looks like a whole studio touched it. Instead, Opus 5.5 built the sequence. That’s fucking ridiculous.

I asked Claude Opus 5.5 to explain how a fusion reactor works by building one you can take apart in your browser. Drag the plasma from 15 million °C, the heat of the Sun's core, to 150 million and watch a city light up. https://airsup.ai/lab

I asked Opus 5.5 again for an engine you can take apart in the browser Pistons against magnets at the same shaft speed. Push the RPM, open the cylinders, slow time down to 1/10 Under 15 minutes, and it did not even move the usage meter on my Max plan. The first experiment took half an hour, this one reused the framework the earlier ones built This is a screen recording, the real thing runs live in your browser Next I want to find where it breaks, so a game is coming

刷到很多用opus 5.5生成的酷炫视频 为啥我用opus 生成的视频满满的像素感 大家都是怎么生成的？

最近用 GPT 6 Astra、Opus 5.5 做了一堆 web 3d 的项目，游戏、室内、Fx、文旅、整活的都有尝试hh 我把这些项目的全部工作流与prompt都整合到了线上的workplace里，之后也会持续更新，感兴趣的朋友可以收藏一波hh

✨审美绝了！Opus 5.5 用 Three.js 把宫崎骏《红猪》经典空战复刻出来了！ 只给了这一句prompt: 「用 web3D 复刻宫崎骏《红猪》日落海面上的经典空战：白绿红涂装战斗机、紫色积云、海面反光、吉卜力运镜，画面写 just breathe.」 建模、配色、云层、尾迹和镜头全部一次对齐，审美直接封神。 #Opus5_5 #宫崎骏 #红猪 #ThreeJS #Web3D #AIGC #vibecoding #审美天花板 #3D #AI #Dario #Anthropic https://x.com/nft_chen/status/2103401780117938683?s=46

I asked Opus 5.5 to generate a realistic 3D model of the Taj Mahal Dome, minarets, marble inlay, even the foundations. One prompt ⤵️ One HTML file.

Claude Opus 5.5 just one-shotted an entire 3D virtual studio in Blender. We used to sell studios like this on Envato. Built in After Effects + Element 3D, each one took our team months. This took one prompt. @claudeai @AnthropicAI @alexalbert__ @bcherny

Claude Opus 5.5 is absurdly good at 3D. I reran one of my recent 3D experiments with Claude Opus 5.5. Same prompt, same references, same real-world scene I had previously built with GPT-6 Astra. The result was much better than I expected. Put simply, for this specific use case, Claude Opus 5.5 was a clear win for me. The biggest differences for me were instruction following and taste. With the same references, Opus 5.5 got much closer to the Zelda Breath of the Wild-style look I was aiming for. The layout felt more intentional. It picked better camera angles. Asset placement made more sense. And I spent much less time fixing basic mistakes. In my Astra version, I had to repeatedly correct things like tree roots not touching the ground, rough geometry being used where proper assets already existed, or objects being placed in ways that just looked wrong. Opus 5.5 got much more of this right on its own. The interesting part for me is that I have almost zero real 3D skills. I mostly give references, look at the result, explain what feels wrong, and keep iterating. With better instruction following and visual judgment, that loop becomes dramatically easier. Still plenty to improve, especially indoors. Next test is whether I can recreate the interior of the house with the same level of quality.

OK so Opus 5.5 SMOKED GPT Astra at 3d modelling this stadium from old plans and reference pics This was a near one shot. 1000x better than Astras one shot (check reply for Astra’s one shot) I gave them the same prompt and the exact same reference plans 👇🏻

Opus 5.5 building a website for a drink brand. -It generated all the svgs -Built the can in 3D -Painted its own labels and then rendered those to make product photo variations All this in a single (albeit, beefy) prompt

Opus 5.5 does a great job in eval of GLSL (OpenGL Shading Language) in threejs. The prompt here is roughly. A fishbowl with reflection, refraction and water caustics. Check it out yourself! And compare it to other models https://threejseval.com/gallery?eval=fishtank

Opus 5.5 turned one sentence into a 3D system that opens in Blender, rigged, textured and ready to animate. It does not render the object. It constructs it, part by part, the way a modeller would if they had a week. > PARSE - the sentence becomes a parts list with real dimensions, nothing implied > MASS - each part built as clean geometry, quads only, no sculpted mush > RIG - joints and pivots placed where the object physically bends > SKIN - materials assigned per surface, scale and roughness set separately > PACK - exported with a named hierarchy, so a human can edit one part Image models stop at the render. What you get is a photograph of a thing that does not exist as a thing. That gap appears the second someone wants to move an arm, swap a material, or send it to a printer. Rigging is what nobody automates. Without pivots you have a statue, and a statue cannot be a system. Topology beats detail every time. A dense mesh looks stunning in a screenshot and folds the moment it deforms. Names carry through the whole export. Every part, joint and material reads the way a person would label it. Studios burn days on modelling and cleanup before anyone gets to the creative part of the job. The prompt format and what each stage returns are laid out underneath.

threejs (@threejs) + Claude Code Opus 5.5 tek prompt ve çıktı