Biological Computing Co. released a text-to-video model on September 22, 2026, saying it runs five times faster and 80 percent cheaper than the base model. It is the company's first commercial product since its founding in February 2026.
The company reported a 19 percent improvement in image stability using its Neural Dynamics Adapter, measured on Minecraft video sequences. That compares with a 15 percent improvement from conventional fine-tuning.
The model is built on an open-source video model and targets AI video generation. Availability begins with early access signups, initially for developers and businesses.
"The adapter holds a pattern that mimics how a signal spreads across the cell culture within limited space," said Alex Ksendzovsky, co-founder and neurosurgeon. The adapter works like a damper that holds neighboring image regions together and keeps disturbances from spreading unchecked.
The announcement follows the company's blog post detailing its approach to turning cell cultures into software modules. The source itself frames the significance as a new method to improve AI inference without biological hardware.
Biological Computing Co. did not say which open-source model it starts from, and whether the biological detour beats purely digital tricks like distillation or caching. The company's real claim is that neurons point to algorithms nobody would have found at a desk.
Source: thedecoder
