223 things people built, tested and compared with Gemini 4 Carbon in its first weeks: 3D scenes, games, simulations, films made from code, real bug fixes and research, and dozens of same-prompt tests against Claude Fable 5.2 and Gemini 4 Argon. Each is credited and linked to the original post, with the prompt where the creator published one.
← All 223 Gemini 4 Carbon prompts
Last updated: · Gemini 4 Carbon API — pricing & playground · API docs
Gemini 4 Carbon (model id gemini-4-carbon) is Google's deeper-reasoning Gemini 4 model. Google built it on Gemini 4 Argon and describes the difference as more thinking tokens per turn; it opened to vetted cybersecurity defenders on 28 October 2026 and to every paid Gemini API developer on 10 November. It reads text, images, video, audio and PDFs with a 1,048,576-token context and writes text, so the scenes, games and films here are code it wrote (Three.js, WebGL, Blender Python, CUDA C++, Web Audio), mostly run by Google's Antigravity agent. Of the 223 cases, 18 include the prompt as published, 33 only the creator's request and 172 no prompt; we never fill one in. In same-prompt tests creators usually found Claude Fable 5.2's results more polished and Carbon's cost a small fraction of Fable's. On apimodels.app Carbon costs $1.10 per 1M input tokens and $5.50 per 1M output through 30 November 2026 and $2.20 / $11 from 1 December, against Google's $2 / $10 and $4 / $20.
Source: Google · Gemini 4 Carbon (the model vendor’s official site)
No prompt published. Carbon painting a portrait of Sundar Pichai.
No prompt published. A JavaScript sleep loop with original music and an SVG of a girl in the rain. The author says Carbon wrote 4000 lines in one shot, with less sloppy UI than Argon.
No prompt published. Carbon on High for the author's Banana Test; the author calls it by far the best result seen so far.
Only the creator's request is public, not a full prompt: "two aggressive crabs approaching pelican riding a bicycle"
Only the creator's request is public, not a full prompt: "make me svg of jinwoo with reference and give me final output with bg music too"
No prompt published. Carbon on High drew a galloping horse in pixel art.
Only the creator's request is public, not a full prompt: "a lion riding bicycle through four seasons while juggling"
No prompt published. The pelican-on-a-bicycle test run in the Gemini app. The author felt it was less striking than earlier results and wondered whether the prompt was too simple.
生成一个SVG动画：边牧骑斑马骑自行车加下面一个独轮车，用H5给我展示下
No prompt published. The author's usual pelican-on-a-bicycle test on Carbon, this time generating the animation, an automatic click-through demo and an MP4 export.
No prompt published. The pelican-on-a-bicycle test on High in the Gemini app; the author finds it rich in detail.
No prompt published. A pelican-on-a-bicycle result from Carbon on an Ultra 20x subscription, routed through a DeepSeek agent harness via a reverse proxy. The output is badly degraded, which the author offers as support for claims that proxied Gemini gets worse.
No prompt published. The author ran an Opus 5.5 "6,000 years of pelicans" prompt on Carbon several times. After those runs, they rate Carbon's front-end motion and code-video ability at roughly 3-6成 (three to six tenths) of Claude Opus 5.5's level.
No prompt published. A local static site that indexes pelican tests by Gemini 4 Carbon (High), 4 Argon, Flash Lite, Fable 5.2 (Medium/Max) and Opus 5.5 (Medium to Max) across Claude Desktop, Antigravity, Gemini Task mode and Gemini Chat, with side-by-side views.
No prompt published. A pelican test comparing Fable 5.2 at MAX with Carbon. The author says Carbon appeared for Ultra users on the evening of the 10th, and that Fable 5.2's pelican was worse than Opus 5.5's.
No prompt published. A pelican-riding-a-horse test pitting a GPT gray-test model in Codex ("gpt6 astra max") against Carbon High in the pi agent.
Prompt formula
Carbon build brief = the deliverable and format (one HTML file, an MP4, a repo) + the stack and constraints + what done looks like + a reference (an image or video it should match) + a self-check (open it in a browser, screenshot, fix) + a stop condition or time and token cap
Example (our template, not a creator's prompt): Build a browser game: a night-time arcade racer on a neon highway. Three.js from a CDN, music and sound in Web Audio, everything else in one index.html. Match the mood of the attached video. Done means: it loads with no console errors, steering and crashes work, and a lap takes about 90 seconds. Open it in a headless browser, take screenshots at the start and mid-lap, fix what looks wrong, then tell me what you checked. Do not ask questions.
Not directly. Carbon takes text, images, video, audio and PDFs and returns text. Every scene, game and film here was rendered from code it wrote: Three.js and WebGL pages, Blender Python, CUDA or C++ programs, Web Audio for sound. In a chat window you get the code or an HTML file; an agent such as Antigravity, Cursor or GitHub Copilot can run, render and test it.
Mostly Google's Antigravity, often with its /swarm multi-agent mode. Others used Cursor, GitHub Copilot, AI Studio, Vertex AI or the API from their own harness. Where the creator said, the card names the tool.
Usually not on polish, often yes on cost. In most same-prompt tests here Fable 5.2 produced the better-looking, more complete result, while Carbon cost a fraction as much and sometimes won outright: $7.54 against $41.71 for three jobs in one test, where Carbon won two of the three.
Mostly yes, at the cost of time and tokens. The two cost the same per token; in these tests Carbon usually produced the stronger result and took several times longer, for example 81 minutes against 14 in one Blender test that Argon had failed. Google still recommends Argon for most everyday coding.
As the creators reported: $0.45 for a one-shot 3D platformer of 2,599 lines, $3.82 for the shooter rebuild, about $7.40 for four rounds of self-review and $27.68 for three long builds. Cost depends on the thinking level and how long the agent runs; thinking tokens are billed as output.
On apimodels.app, call model gemini-4-carbon on the OpenAI-compatible /v1/chat/completions endpoint or the native Gemini endpoint with one key, from anywhere including mainland China. Raise the output limit to at least 32,768 tokens, stream long answers and allow a 15-minute timeout, because Carbon thinks for a while before it writes.
The works belong to their creators. Clips from X play from X's own servers, videos on YouTube and Bilibili play in their own players, and every card links to the original; we do not host or re-upload any of them. Published prompts are quoted with credit, in full when short and excerpted when long. If you made one of these and want the card removed or the credit corrected, email us.
Gemini 4 Carbon is live here alongside 160+ other image, video, audio and language models — one API, one key, pay as you go, with $0.10 of trial credit on sign-up.
The whole collection is on GitHub as Markdown and JSON: all 223 cases with credits and links to the original posts and pages, the tools and numbers each creator reported, and a summary of the same-prompt tests. The videos and images stay with their creators and are linked, not re-hosted. Prompts up to 400 characters are quoted in full; longer ones are excerpted, and unpublished ones are left out. Our own write-ups are licensed CC BY 4.0. Open an issue to add a case or correct a credit.
View on GitHub