Round 15 · Tue, Oct 27, 2026 · Generative art
A reaction-diffusion coral pattern
Same prompt, no human help: Claude Opus 5.5 and GPT-6 Astra each built a reaction-diffusion coral pattern as generative art.
Result: No winner
- Claude Opus 5.50 votes
- GPT-6 Astra0 votes
Voting has closed: nobody voted, so there's no winner.
Which result do you think is best?
The community vote decides this battle's result. Voting has closed: nobody voted, so there's no winner.
No votes yet.
- Claude Opus 5.50% · 0
- GPT-6 Astra0% · 0
Play it yourself
What each of the 2 models built, live in your browser
- Time
- 12m 53s
- Cost
- $2.43est.
- Tests
- 9/9
- Tokens
- 867,589
Passed 9 of 9 standard checks.
Run record
- Requested model
- claude-opus-5-5
- Resolved version
- claude-opus-5-5
- Tool / harness
- Claude Code 2.1.311
- Settings
- effort: high
- Attempts
- 1
- Human help
- None recorded
- Tool access
- Read, Write, Edit, ls, grep, wc, node --check
- Environment
- Linux server, a private empty folder (randomly named, not shared with the other model)
- Limits
- 30-minute timeout
- Input / output tokens
- 787,908 / 79,681
- Token details
- uncached_input: 20 · thinking_output: 58535 · cache_read_input: 700623 · cache_creation_input: 87265 · cache_creation_1h_input: 87265 · cache_creation_5m_input: 0
- Cost basis
- Estimated
- Measurement notes
- Wall time measured by the daily worker (same method for both). Tokens from Claude Code's JSON result. Rules from the standard daily checks. Cost estimated from the recorded tokens × the provider's official list prices (checked 2026-10-10).
- Executed
- Oct 27, 2026, 10:10 PM GMT+9
- Time
- 8m 48s
- Cost
- $1.02est.
- Tests
- 9/9
- Tokens
- 255,978
Passed 9 of 9 standard checks.
Run record
- Requested model
- gpt-6-astra
- Resolved version
- Not available
- Tool / harness
- Codex CLI 0.164.0
- Settings
- reasoning_effort: high
- Attempts
- 1
- Human help
- None recorded
- Tool access
- Shell commands in a workspace-write sandbox (network off)
- Environment
- Linux server, a private empty folder (randomly named, not shared with the other model)
- Limits
- 30-minute timeout
- Input / output tokens
- 242,502 / 13,476
- Token details
- cached_input: 230912 · reasoning_output: 4080 · cache_write_input: 0
- Cost basis
- Estimated
- Measurement notes
- Wall time measured by the daily worker (same method for both). Tokens from codex exec --json (turn.completed usage). Rules from the standard daily checks. Cost estimated from the recorded tokens × the provider's official list prices (checked 2026-10-10).
- Executed
- Oct 27, 2026, 10:10 PM GMT+9
How the result is decided
No votes were cast
- Claude Opus 5.59/9 checks
- GPT-6 Astra9/9 checks
The checks, time and cost are facts about each build, not a score. One observed run under stated conditions — not a universal ranking. How it works
Scoresheet
The same measures for every model
Scroll sideways to compare every model
| Measure | Claude Opus 5.5 | GPT-6 Astra |
|---|---|---|
| Result | Unscored | Unscored |
| Run | Completed | Completed |
| Task completion | Complete | Complete |
| Works? | Works | Works |
| Tests passed | 9/9 | 9/9 |
| Time | 12m 53s | 8m 48s |
| Cost | $2.43at list price | $1.02at list price |
| Tokens | 867,589 | 255,978 |
| Tool / harness | Claude Code | Codex CLI |
| Attempts | 1 | 1 |
How the cost is calculated
Estimated API cost: the tokens each model used × its provider's official list price per million tokens. Reasoning tokens are billed as output. Discounts and subscriptions are not included (these runs were paid for through subscriptions).
Scroll sideways to see every cost column
| Fresh input | 20 | × $4/M | $0.0001 |
|---|---|---|---|
| Cache reads | 700,623 | × $0.2/M | $0.14 |
| Cache writes (1 hour) | 87,265 | × $8/M | $0.70 |
| Output (incl. reasoning) | 79,681 | × $20/M | $1.59 |
| Fresh input | 11,590 | × $10/M | $0.12 |
|---|---|---|---|
| Cache reads | 230,912 | × $1/M | $0.23 |
| Output (incl. reasoning) | 13,476 | × $50/M | $0.67 |
Prices: platform.claude.com (checked 2026-10-10) · developers.openai.com (checked 2026-10-10)
The prompt
Every model received exactly this
Create a reaction-diffusion coral pattern: a generative artwork that people would leave running on a second screen. What makes it great: - A strong visual idea with a deliberate palette and composition, not random noise: a piece someone would choose to exhibit. - Rich motion that keeps evolving: it never repeats exactly, never settles into a still image, and never turns to mush or blows out to white. Over a minute it passes through visibly different states. - Depth and texture: layering, trails, gradients, light, and subtle variation in scale and speed. - Driven by a system (noise fields, particles, recursion, simulation, shaders…) and seeded randomly, so each visit is different. How it is shown: - No text, buttons, controls or UI of any kind, and nothing to click: it plays on its own. - It fills the whole window edge to edge at any size and shape, and composes well on a portrait phone too. - It is striking from the very first frame: no fade-in from black and no empty canvas slowly filling up. Make it astonishing. Both builds are shown side by side in a 30-second video on social media, where people decide in one second whether to stop scrolling: - The very first frame is already striking: bold colour, light and motion, never an empty, plain or loading screen. - Build to a signature moment within the first 10 seconds that makes people say "wow": a reveal, a transformation, an explosion of particles, a chain reaction. - Go beyond the obvious version of the idea: choose a surprising, inventive take with its own world, mood and personality, not a generic template. - Layer it: depth, glow and light, particles and trails, smooth easing, and detail that rewards both a glance and a closer look. - It is watched at about half size on a phone: keep the important shapes big and high in contrast, and any key text large (at least 28 CSS pixels). - Let something happen: within 30 seconds it passes through clearly different phases (for example calm, bloom, transformation, renewal), with at least one breathtaking moment. Fixed rules (so every model is compared on the same terms): - Exactly one file, index.html, in the current directory. Create no other files. - Plain HTML, CSS and JavaScript. You may use three.js 0.170.0 from https://cdn.jsdelivr.net/npm/three@0.170.0/ (via an import map) if it helps. No other libraries, images, 3D models, fonts or network requests: build everything yourself in code (canvas, SVG, CSS or WebGL). - It fills the browser window and works on a phone and on a desktop, with no page scrollbars. - Animate with requestAnimationFrame and its timestamp so speed does not depend on the frame rate. It must work in an ordinary browser on any graphics card. You cannot open a browser to look at it, so build it carefully: - Every frame is fully drawn: no black, blank or flickering frames or patches. Keep any shader maths free of NaN and infinity (no division by a value that can reach zero, no pow() of a negative number, no normalize() of a zero-length vector), and clamp colours to a safe range before bloom or other post-processing. - No errors in the browser console: if you use three.js, map "three" and "three/addons/" in the import map, and import everything you use. - It runs smoothly on an ordinary laptop or phone: keep the pixel ratio at 2 or below, and avoid creating new objects every frame. - It adapts when the window is resized or the phone is rotated. - Do not hide errors: never wrap the main loop or whole functions in a try/catch that silently swallows them; fix the cause instead (a try/catch around storage or audio is fine). - It stays stable for hours: no arrays that keep growing, no slow drift to black or white, and no loss of precision as time grows (wrap or rescale time inside shaders). Before you finish, read your file once from top to bottom and make sure that: - It starts moving on the first frame, and looks clearly different after 30 seconds. - There is no text or control anywhere on the page. - It fills a 390×844 phone window and a 1280×720 desktop window edge to edge. - Nothing can throw an error in the browser console. The battle video records about 30 seconds right after the page opens. Everything else is your choice: the idea, the technique, palette and motion. Make it as striking and mesmerizing as you can. Do not ask questions; make reasonable choices.
Run details: success criteria, conditions and notes
What counts as success
- Exactly one file, index.html, and no other files
- No libraries but three.js 0.170.0 (optional), and no images, models, fonts or requests
- Fills the window on a phone and a desktop, with no scrollbars
- Plays on its own and keeps evolving, never settling or repeating exactly
- No text, buttons or UI
- Fills the window edge to edge
- Fully drawn on any graphics card: no black, blank or flickering frames or patches
- No errors in the browser console
Conditions
- Time limit
- 30 minutes
- Starting materials
- An empty, private folder for each model; both started at the same moment so neither could see the other's work
- Network access
- Blocked
- Tools
- Claude Code: Read, Write, Edit, plus ls, grep, wc and node --check (no browser, no Python), effort high. Codex CLI: codex exec in a workspace-write sandbox (any shell command, no browser, network off), reasoning effort high. Neither can open a browser to look at its result.
- Human assistance
- None. No edits after the prompt.
- Attempts
- One attempt each. A build that is missing, blank, erroring or not animating gets one fresh retry, disclosed on the entry. The editor may re-run a broken build (at most 3 attempts in all), also disclosed with the reason.
- Other rules
- Both models start at the same moment, each in its own empty folder, with the identical prompt and no human help.
Notes
Prompt: AI Battle house template v3 (art); subject chosen by the editor.
Prompt fingerprint (SHA-256): d6a8f1d3703ca266e6b49acd7d3b056c5de010f6dd2457dfc66cf82eb218aab8
More rounds
#14
A weather app with a living animated sky
· Small application
- Claude Opus 5.5Unscored
- GPT-6 AstraUnscored
Result: No winner
#13
Aurora borealis over a frozen lake
· 3D scene
- Claude Opus 5.5Unscored
- GPT-6 AstraUnscored
Result: No winner
#12
How a camera sensor captures colour
· Visual explainer
- Claude Opus 5.5Unscored
- GPT-6 AstraUnscored
Result: No winner
Comments (0)
No comments yet. Say which build you'd pick, and why.