"Luciferus": An Uncensored AI Service Sold on Exploit Forum Writes a Working RAT on Its Cheapest Tier

Cybercriminal buyers, as a malware development service, and indirectly anyone they target

Sophos Counter Threat Unit reported on September 15, 2026 on Luciferus, a subscription AI service advertised on August 24 on the Exploit forum by a persona named "Optimus_Prime" as "an AI system that responds to requests without moral or ethical constraints" and positioned as an alternative to jailbreaking ChatGPT and Claude. It claims a proprietary 120-billion-parameter model; Sophos assessed with low confidence that it is built on Alibaba's Qwen. Forum tiers were Inquisitor ($35), Archdevil ($55), and Prince of Darkness ($75), with public site tiers from $22 to $47.14 and a custom "Individual Embodiment" deployment. Asked for a simple remote access trojan in Python, the Junior tier returned a Russian-language explanation of the networking and command execution functions followed by source code. Sophos could not verify the architecture, performance, or privacy claims

Not a vulnerability. Open-weight models fine-tuned without refusals and resold as a service remove the need to defeat a commercial model's safeguards at all