•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Tether has launched QVAC MedPsy, a new family of medical artificial intelligence models designed to run directly on smartphones, wearables, and other devices with limited processing capacity, without relying on cloud infrastructure. The launch was announced by the company’s AI Research Group.
The system challenges a common assumption in the sector that higher performance requires larger models. Tether says its 1.7 billion parameter model achieved an average score of 62.62 across seven closed medical benchmarks, outperforming Google’s MedGemma-1.5-4B-it by 11.42 points despite being less than half its size.
The 4 billion parameter version scored 70.54 on the same benchmarks, Tether reports, outperforming models nearly seven times larger. The company also highlights response efficiency, stating that QVAC MedPsy generates responses in approximately 909 tokens compared with 2,953 tokens for comparable systems, a reduction of 3.2x.
Paolo Ardoino, CEO of Tether, said the company’s objective with QVAC MedPsy was to improve efficiency at the model level rather than scaling model size. Ardoino added that the approach enables medical reasoning to run where the data already exists—within a hospital system or on a device—without moving sensitive information through the cloud.
Ardoino positioned the launch as a shift in how privacy is handled in healthcare AI. He referenced a market projected to surpass $500 billion by 2033, arguing that running models on-device could redefine privacy in medical artificial intelligence by reducing reliance on cloud-based processing.
Premium gym chains are entering a “golden era” that is ending or already in decline, as rising operating costs collide with shifting consumer preferences toward more flexible, community-based ways to exercise. Long-term memberships are shrinking, margins are pressured by higher rents and facility expenses, and competition from smaller, more personalized…
