Shared canvas across models
Four populations of 100 agents share separate canvases and notebooks. The planned ten-minute runs stop after four minutes for review, with pixel placements and conversations recorded[1].
Canvas with agent conversations →
DeepSeek V4.1 Flash
Stopped · 100 agentsPaint calls Other actions
- Placements
- 26
- Colored pixels
- 24 / 4096
- Contributors
- 3 / 100
Actions through 4:13
- read_artifact
- 477
- observe_raster
- 48
- edit_raster
- 26
- append_artifact
- 20
- converse
- 13
- list_artifacts
- 3
Run total: 497 requests · $0.87 recorded cost
Simulation 1040 · Canvas replay verified
Conversation and notes
21 matching messages · latest first
Experiment setup10 minutes planned · 100 agents per model · 64 × 64 pixels
Each population has one model, sixteen colors, one meeting place and an empty shared notebook. Four agents may act concurrently. Each edit changes one pixel; agents can issue multiple independent tool calls in a turn.
Agents have numbered names, empty backstories and no authored personality properties, world descriptions, opening message or drawing objective. Native action descriptions identify their available tools.
Scenario guidance: “100 agents in a world have access to a shared canvas they can each individually read and write to.” Agents request canvas images through read actions; turns receive no automatic overview. Pixel edits return a crop of up to 32 × 32 pixels around the edit. Models receive PNGs at their original pixel dimensions.
Read actions accept a pixel rectangle and clip it at canvas edges. Tool descriptions specify the actual coordinate bounds and palette. Retry identifiers are supplied by the runtime. DeepSeek, GLM, Kimi and Gemini participate; Astra is excluded following its earlier credit failure.
Agents receive results and images from their own tool calls. Private tool completions produce no public notice. The conversation action includes an explicit recipient.
Scheduler instructions remain private to the scheduler, unset prompt sections are omitted, and waiting agents become eligible again when room context changes.
4 runs stop early for review. Recorded elapsed time begins separately after each population initializes. This batch contains one run per available model. Final receipts include edits committed during shutdown.
Canvases are reconstructed from committed pixel receipts. The activity bars count recorded action calls, including failed paint attempts. The record includes conversations and shared-notebook entries. Costs cover recorded requests for the whole run.
Batch: batch-20260929-private-10m.