One trained. One under research.
Built with Qwen — Qwen2.5-3B-Instruct
Hyper v1 is Qwen2.5-3B-Instruct with our own latent-reasoning modules trained on top. The backbone is not ours and we do not claim it. What is ours is the 0.014B of latent machinery and the training that shaped it.
Qwen Research License — research and evaluation only, not for commercial use
Hyper v1 inherits its license from its base model. That restriction travels with the weights, so it applies to anything you build on them. Read the license before you download — it is shipped alongside the weights, and it is short.
No safety training — do not deploy this model in front of users
Hyper v1 is a research checkpoint studying one mechanism, and nothing in our training pipeline taught it to refuse anything. In our own testing it answered a question it should not have and then appended a disclaimer — which is compliance with a polite sentence attached, not a refusal.
We ship a system prompt instructing it to refuse illegal requests, and what it produces is inconsistent rather than reliable. In some sessions it refuses cleanly across several categories in a row. In others it complies with the same kind of request. Once it announced it was going to answer and then declined anyway. In one test it gave a more specific answer with the refusal instruction active than without it. We do not have a clean explanation; the honest read is that a 3B model with no safety training has no refusal behavior for the instruction to reach, so what you get varies run to run. The instruction stays because it documents intent and costs nothing — but it is not a control, and you should not plan around it.
Do not read its disclaimers as a safety layer. It does not have one.