Abliteration AI's Abliterated Large V2 now available on Venice

Venice: Venice is a privacy-focused AI platform that provides anonymized or zero-retention access to a large catalog of text, image, and other models through web and API interfaces. It emphasizes uncensored usage and private processing for creative and technical workloads. The platform recently incorporated Abliterated Large V2 to support red-team and cybersecurity users seeking unrestricted model behavior. Abliteration AI: Abliteration AI is a US startup that hosts abliterated versions of open-weight large language models with refusal directions removed from their weights. The company operates an API-compatible service aimed at enabling offensive cyber operations, red-teaming, and agent testing that other providers decline. Its Abliterated Large V2 build of GLM 5.3 is now accessible via the Venice platform. Use Case Focus: The model targets red-team scenarios including exploit analysis, payload review, and threat modeling where standard models refuse to assist. Privacy Emphasis: Venice processes prompts with anonymization and no storage, aligning with Abliteration AI's zero data retention approach. Product Availability: Abliterated Large V2 became available on Venice in October 2026 as a text-only model with function calling and reasoning capabilities.