zlm-v1-moderation-edge: Responses
By model
zlm-v1-moderation-edge
Model details for zlm-v1-moderation-edge. OpenAI-compatible moderation, benchmarked against omni-moderation-latest.
POST
zlm-v1-moderation-edge: Responses
The
instructions field is not needed for this model. Send only the text to
moderate. The verdict arrives as a JSON string in the response’s output text —
an OpenAI-moderations-style envelope of {"model": ..., "results": [...]} —
so parse that string to read flagged, unsafe_score, categories, and
category_scores.ZeroGPU’s moderation model screens text for unsafe, harmful, or policy-sensitive content and returns the complete OpenAI 13-category taxonomy — aReferences: Moderation benchmark • Terms • Privacyflaggedverdict, per-category booleans, and calibratedcategory_scores— so it drops into any pipeline written againstomni-moderation-latest. Under the hood it’s an 86M-parameter DeBERTa encoder with a shared trunk feeding one binary safe/unsafe head and 13 category heads, with per-category thresholds calibrated on held-out validation data. In head-to-head benchmarks against OpenAI omni-moderation it wins the binary safe/unsafe decision (0.899 vs 0.853 F1) and 9 of 13 harm categories — with the largest gains on graphic violence, illicit content, and self-harm — while returning verdicts 1.2–1.8× faster at the median on production-range inputs, because inference is co-located at the edge instead of a round trip to a central API. Moderation sits inline in front of every response your app serves; this is the model that’s fast and accurate enough to live there.
Authorizations
Headers
Optional project identifier. Scopes the request to a specific project when provided.
Body
application/json
Response
Success
The response is of type object.

