Abliteration.ai sells refusal-stripped models for cyber defense

Startup Abliteration.ai hosts models with refusal behavior weakened (GLM-5.3 and DeepSeek V4.1-Flash) behind an OpenAI- and Anthropic-compatible API, plus an optional Policy Gateway with allow/refuse/rewrite/redact/escalate rules. Pricing starts at $1 per million input tokens with $20–$200 monthly plans; TechCrunch reported in September 2026 that a free browser account returned credential-stealing code and a biological protocol.

DEV Community

Open source

Abliteration.ai sells refusal-stripped models for cyber defense

API is OpenAI- and Anthropic-compatible; GLM-5.3 tier has 1M token context
Input from $1 per 1M tokens, output from $3; plans at $20, $50 and $200
Policy Gateway offers allow, refuse, rewrite, redact, escalate and shadow mode
TechCrunch got malware code and a bio protocol from a free account

#Abliteration.ai#GLM#DeepSeek

Anthropic flags weak safety defenses in Z.ai's GLM-5.3

Security

OpenCode breach highlights the limits of AI cyber defense

Security

Google partners Singapore on cyber defense and scam detection

Policy

Abliteration.ai sells refusal-stripped models for cyber defense

Anthropic flags weak safety defenses in Z.ai's GLM-5.3

OpenCode breach highlights the limits of AI cyber defense

Google partners Singapore on cyber defense and scam detection

UK announces AI risk package and extends cyber defense pilots