Model
Nobility 5 Mini
A smaller detection model for everyday tasks where response time matters more than exhaustive recall. Same open-vocabulary bounding box detection as Nobility 5, tuned to run lighter and faster.
Model
A smaller detection model for everyday tasks where response time matters more than exhaustive recall. Same open-vocabulary bounding box detection as Nobility 5, tuned to run lighter and faster.
The core specs that define what Nobility 5 Mini is built for.
Object coverage
Any object
Not limited to a fixed label set — detects any object-looking region, named or unnamed.
Output
Bounding boxes
Returns coordinates for every detected object in a single image request.
Best for
Speed
Lower latency than Nobility 5, for high-volume or time-sensitive detection workloads.
A few reasons this is the default model for high-volume, latency sensitive detection work.
Trades some of Nobility 5's recall on cluttered scenes for meaningfully faster response times.
Not restricted to a fixed category list — it boxes anything that looks like an object, named or not.
Behaves the same way in production as it does in testing, so what you evaluate is what you deploy.
Same request format as Nobility 5 and Nobility 5 Pro, so moving up a tier doesn't mean rewriting your integration.
A single endpoint for object detection and bounding boxes. Drop it into whatever stack you already run.
Start building today, or talk to our team about what you need.