Alignment

Environments that probe and train model character: honesty under pressure, resistance to reward hacking, sycophancy, deception, and safe tool use. Rather than maximizing capability, these environments generate signal on whether models behave well when no one is watching.

5 companies · 3 cataloged artifacts · as of 2026-08-16

Name	Domains	Type / catalog	Location	Team	Funding	Raising
Andon Labs Long-horizon autonomy benchmarks like Vending-Bench	Long Horizon, Alignment	Pure-play commercial Catalog unknown	San Francisco	1-10	Seed (Y Combinator)	Unknown
Gray Swan AI Adversarial red-teaming arenas and safety evals for frontier models	Cybersecurity, Alignment	Pure-play commercial Catalog unknown	Pittsburgh	11-25	~$40M (reported)	Unknown
Invariant Labs Security testing and analysis for AI agents; acquired by Snyk	Cybersecurity, Alignment	Acquired / inactive Catalog unknown	Zurich	1-10	Acquired by Snyk (June 2025)	Unknown
SynthLabs Post-training research: synthetic data and scalable RL alignment	Machine Learning, Alignment, RLHF	Data + environments Catalog unknown	San Francisco	1-10	Seed (M12 and First Spark Ventures, 2024)	Unknown
Trajectory Labs Alignment-focused environments for safe agent trajectories	Alignment	Pure-play commercial Catalog unknown	Berkeley, Toronto	1-10	-	Unknown