Threat Pulse 62High
Theory: abliterated open-weight models (DeepSeek, GLM, Kimi) are powering the attack wave
Severity 3Status unknownreportedAI and software supply chainAI-related

Flare's analysis found dark-web mentions of abliterated DeepSeek, GLM and Kimi models up 745 percent in 2026, with vendors selling guardrail-free access and seven AI-driven attack cases from May to October. Anthropic assessed GLM-5.3 as near-Mythos in cyber capability, and Palo Alto startup Abliteration launched a cheaper unrestricted model based on DeepSeek-V4.1-Flash on October 23. Critics note most sold access is resold API keys, not true weight edits.

What to do: Open question; Anthropic and Flare both expect more.

AI angle: This is the 'obliterated DeepSeek' theory Joel heard; the strongest sourced version points at GLM-5.3 as much as DeepSeek.

· 16 days ago · criminal AI marketplaces · Security Boulevard / Flare
SourcesSecurity Boulevard / FlareGizmodo (Abliteration launch)PlayCISO
On X
@skeptruneunverified

DEEPSEEK ABLITERATED REFUSES NOTHING & IS JUST AS INTELLIGENT AS THE REGULAR MODEL!! that should scare the hell out of you.

Open on X
@YishaiBackunverified

argues for

Open on X
@mpfethunverified

argues against

Open on X
Why this score

Supported by market data and a few documented attacks, but no evidence yet ties a specific abliterated model to PHP2Shell or the big October breaches.

Back to the board

Updated Nov 5, 01:45 UTC. Refreshes on its own.

Add a threat

Paste a link to an advisory, an article or an X post. A person reads every submission before it shows; nothing posts on its own.

Share

The picture people see when the link is posted
Post on XSave picture

This picture is what X, iMessage, Slack and Facebook show under the link. It is drawn fresh from the live board: the board and chart pictures update every 10 minutes, an event picture follows that event.