Laya
Small encoder models that answer routing and triage questions in one pass
Emerged finds new products worth a look, daily — what’s new today
W
If your app asks an LLM to classify text and then parses the string, this returns a calibrated probability per option from a 726 MB model you can run offline.
wattai is a decision model distilled from SmolLM2-360M. You give it text and a list of choices; it scores each choice and returns a probability distribution rather than generated text, so there is nothing to parse. It publishes benchmarks, including 96.3% on CLINC150 and its calibration error, and is candid about limits: a probability says which option fits best, not whether any fits. Weights are Apache-2.0, and a hosted API allows 1,000 keyless calls a day during the preview.
More like this
jev-align
Teach a Jev classifier your labels by reviewing only its least certain answers
SemIf
Read a yes/no probability straight out of a small open model, no generated text
Kev
Open 0.8B to 9B decision models for choice, yes/no and rating questions
Flide
Exact, shared-context inference server for small Jev-style decision models
L
Laya Vision
A 256M model that answers typed questions about an image with calibrated odds