Observed arrival · 2026-11-05

Abliterate: an API for refusal-removed models

abliterate.tech Visit website

Editorial interest 76/100 Selection signal · not a rating of the site

A pre-launch API offering open-weight models with refusal behavior removed, aimed at security testing, red-teaming, and research.

Landing page captured for the 2026-11-05 issue.

For: Security testers, red-teamers, and model researchers
Worth noticing: Its benchmark panel reports that the 1.7B model refused 0 of 100 harmful-behaviour prompts; models are not kept running.

Field notes

The page shows Python, Node, and curl examples targeting an OpenAI-compatible endpoint, with links to a quickstart, API reference, errors, and console documentation. Its benchmark panel reports 0 of 100 harmful-behaviour prompts refused by the 1.7B model, alongside a reported 163–219-second first-request delay when no GPU is running. The site labels itself a pre-launch preview: payments are in test mode and models are not kept running.

Observed signals

Read the marks

Editorial observations of this landing page, not a rating.

○OpenPublic substance visible

⊠LoginAccess appeared gated

$PaidCommerce or pricing visible

✦PrettyNotable craft visible

◎NicheUnusually specific use

One card from the complete issue

The Rescue Mosaic

365,038 arrived 1,000 judged 978 catalogued Enter the complete issue