Abliteration AI launches Large V2 model based on GLM-5.3
Abliteration AI has launched abliterated-model-large-v2, a model based on GLM-5.3 with further modifications and fine-tuning, offering 1 million token contexts and three inference modes.
Abliteration AI Launchedabliterated-model-large-v2Its official documentation states that the model is based on the open-weight GLM-5.3 and has been further modified and fine-tuned.The model is in plain text format, provides 1 million token contexts, and supports three inference modes: low, high, and max.
This service is provided through Abliteration AI's OpenAI-compatible API, supporting streaming output, tool calls, and structured output. The vendor positions it as a model to reduce rejection restrictions; however, such weight modifications may weaken the safety margins of the original model, and the official capabilities and safety claims lack independent verification.
refer to:Abliteration AI Model Documentation