Benchmarks

The current evaluator uses ten task templates with ten fixture variants each. The v1 and v2 agents are deterministic scripted controls with explicit tool requests. A perfect control score verifies the fixture path; it does not establish autonomous AI capability.

Generated 30 September 2026. Groundedness, hallucination rate, tokens and monetary cost are unobserved and remain null. No unsupported numerical estimate is substituted.

Inspect the full report → · Download JSON · Reproduce from source

Project integration checks

Five local contract checks passed: Personizer workflow/permissions, NovaDB isolation and durable reopen, Genopedia lexical provenance, Neurapedia metric calculations, and eight robotics control/fault scenarios. These checks do not measure clinical, scientific or hardware performance.

Inspect integration evidence →