What it sells
Training gyms for computer use and software engineering work
Business model breakdown
Buyer: who owns the budget?
UnverifiedAI research/post-training teams at frontier labs (and enterprises) buying evals and RL training environments.
Trigger: what makes them buy now?
UnverifiedLabs need realistic, verifiable computer-use and coding environments to benchmark and post-train agents with RL.
Workflow: what inputs become what output?
Build high-fidelity clones of software suites (EHRs, enterprise apps), 3D/Blender simulation and coding task environments with automatic verifiers; domain experts write tasks; run graded model runs across Llama, Claude, Gemini, GPT with their harness.
Economics: what it saves and how it charges
UnverifiedPer-engagement delivery of environments, evals and graded runs (project-based services/data revenue); no figures published.
Distribution: how it reaches customers
Founder-led engagements with frontier labs ("Get in Touch"), benchmarks (BlenderBench, SWE-Marathon) as credibility marketing.
Funding history and coverage
| Round date | Stage | Amount | Investors | Source |
|---|---|---|---|---|
| — | Y Combinator Spring 2025 | — | Y Combinator | ycombinator.com |
Profiles summarize public reporting and company websites. Business-model notes are MaxBC's reading of that material, not confirmed facts; unknowns are left blank rather than guessed. Updated 2026-10-12.