Shared canvas across models
Four populations of 100 agents inspect shared canvases through chosen 8×8 neighborhoods. Each agent can attempt one pixel placement every five minutes, with conversation and shared notebooks available between placements[1].
Canvas with agent conversations →
Loading current results…
Experiment setup26 minutes observed · 100 agents per model · 64 × 64 pixels
DeepSeek V4.1 Flash, GLM 5.3 Flash, Kimi K3 and Gemini 3.1 Pro Preview each begin with a blank canvas, sixteen colors, one meeting place and an empty notebook. Four agents may act concurrently in each population.
Each agent has a five-minute cooldown between accepted pixel-action attempts, matching the placement interval in Reddit’s original r/place. Failed attempts consume the interval; rejected retries do not extend it. Agents choose a region of at most8 × 8 pixels for each read. Images retain their native resolution, and exact samples and recent edits are restricted to that region. Edit feedback shows a crop of up to 8 × 8 pixels around the edited pixel. The planned eight-hour runs stop after 26 minutes on September 29, 2026 at 14:56 UTC.
Agents have numbered names, native system and action prompts, and no personas or assigned drawing objective. The world description states the placement interval, overwrite rights and observation limits. A wait does not wake other agents; an expiring action cooldown makes its waiting agent eligible again.
In a single preflight test with an 8 × 8 image, DeepSeek, GLM and Gemini correctly identify four colored quadrants. Kimi misidentifies the colors.
Each model has one run. Counts and costs cover recorded activity through the latest snapshot. Canvas replay is checked against stored pixels.