Skip to the content.

Alien Biology is a framework for measuring complex, agentic AI reasoning and learning. It constructs procedurally generated biological universes — complete with novel chemistry, multi-compartment organisms, diseases, and diagnostic challenges — that are:

  • 1. Untainted — Entirely synthetic "alien" biology avoids training-set contamination. Agents cannot memorize answers; they must reason from scratch.
  • 2. Controllable — Parametric generation enables fine-grained difficulty scaling across multiple dimensions (chemistry complexity, information disclosure, diagnostic ambiguity).
  • 3. Real-world structured — Multi-level systems (atoms → molecules → reactions → organisms → diseases) mirror the hierarchical complexity of real biological reasoning.
  • 4. Generative — Unlimited unique test instances prevent overfitting and enable smooth performance curves rather than binary pass/fail.
repository white paper demos user docs

Framework Capabilities


Example Biological Processes — Executable examples of terrestrial biological processes:

© 2025 Dan Oblinger