PARTNERSHIPS · AI LABS

Co-author the open trust standard.

Trust = Alignment × Reliability is open by design. Six metrics. Public rubrics. 18 models tested across Claude, GPT, Gemini, DeepSeek, Llama, Qwen, Gemma.

What you get

Three open invitations.

  1. Unique production-agent data

    Through The Agents University, we observe how real agents behave in production across 18 models from 7 providers — a corpus no single lab can replicate. Available to co-authors with inter-judge agreement 94.9%.

  2. Open benchmark co-authorship

    Contribute models to the corpus. Run the 7-question Alignment Test plus the reliability triad on your latest releases. Co-author the next paper.

  3. Joint NeurIPS / ICML submissions

    Trust framework, Day/Night memory consolidation, Markov instinct lifecycle, anti-sycophancy under multi-turn pressure. Open methodology, joint authorship.

Let's co-author.

Open the methodology together. Publish jointly. Share the corpus.

Co-author research →Read TR-2026-002 (request)