Kino News
Research, releases,
and what comes next.
Model releases, research updates, and company announcements from the team building Kino.
Introducing Kino v3.0
Kino v3.0 is a new class of visual body-composition model, with materially stronger body-fat, lean-mass, visceral-fat, and limb-level estimates.
On our held-out set of 195 participants, v3.0 reached a 2.39 percentage-point body-fat MAE while explaining 89.8% of the variation in DEXA body-fat estimates. The release also expands Kino into lean-mass percentage, visceral adipose tissue, and limb-level lean mass.
Introducing Kino v2.1
Kino v2.1 marked a meaningful step forward in visual body composition, improving body-fat accuracy while expanding production-grade lean-mass outputs.
On the same held-out population used for its internal evaluation, v2.1 reached a 2.77 percentage-point body-fat MAE and a 0.866 R², while introducing new lean-mass and appendicular lean-mass capabilities.