OpenAI CSAM Classifier - Online CSAM Precision: leaderboard
Publisher evaluation of one released CLIP+MLP image classifier. The two output signals belong to the same checkpoint. Held-out images were randomly split and perceptually deduplicated; related images can remain across splits. Native fractions, higher is better. Precision on sampled ChatGPT traffic at the default operating threshold; traffic prevalence differs from the held-out set. No absolute online recall or sample count is disclosed. Default threshold for likely-csam: 0.996324. Metric: Precision (0–1). Source: https://deploymentsafety.openai.com/csam-image-classifier/csam-image-classifier.pdf (table 7, page 13).
Metric: Precision (0–1). Source: deploymentsafety.openai.com. 1 model tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | CSAM Image Classifier | 0.298 |
Interactive version: theaggregate.ai/benchmark?slug=openai-csam-classifier-online-csam-precision · How It Works · Data refreshed daily, snapshot 2026-11-17.