Quick Takeaways:
- Vetology has rebuilt and re-released all 94 AI radiology classifiers on a new model architecture, with the updated models rolled out across its live library between June and August 2026.
- The new architecture is designed to interpret global information within individual images and across linked images in a study, more closely reflecting how veterinary radiologists assess cases. Vetology says each classifier met its updated performance threshold before release.
- The company has published condition-level performance data for all classifiers and continues to invite independent researchers to evaluate its models. AI reports are intended to support, rather than replace, veterinarians’ clinical judgment, with complex cases available for review by Vetology’s board-certified radiologists.
Why It Matters?
Vetology’s rebuild of all 94 AI radiology classifiers on a new architecture that interprets both single images and linked studies more like a human radiologist strengthens its core value proposition of fast, consistent AI screening backed by board‑certified teleradiology. By publishing condition‑level performance data and inviting independent evaluation, Vetology is leaning into the transparency and post‑deployment monitoring that professional bodies have highlighted as essential for trusted veterinary AI, which can help it win and retain clinic customers in a competitive teleradiology/AI market where practices increasingly expect AI to support, not replace, clinical judgment.
Source: PRnewswire
