Questions & answers

Kvantim LLM is a classic transformer model rebuilt from an open-source model base so it can keep being trained and specialised for a domain, instead of being a fixed snapshot you only prompt around. Before release we will name the base model and licence, state what we changed, and describe the method — continued training, fine-tuning or adapters — in plain terms. It is in development.

Kvantia Harness — the AI-worker supervision layer — is in early access on the web (private demo), with Windows and macOS apps in development. Kvantia AI runs as a working prototype. Kvantim LLM is in development. The roadmap on the home page shows the sequence, with one real status per item.

The next milestones: deeper product work on Harness, model training and evaluation for Kvantim LLM, early access for Kvantia AI, plus the team, licensed domain data and compute to do it. This is deep product and model work and carries real risk — the invest page sets out the evidence, milestones and risks.

Join the early-access waitlist to be first into each product as it opens. To invest or pilot a model, use the invest page or contact us and we'll follow up with the deck or an early-access plan.