Abliterated Model Large is the larger sibling in abliteration.ai's lineup, described by the provider as derived from the open-weight GLM-5.2 model and further abliterated and fine-tuned by Abliteration AI. Where the standard Abliterated Model is multimodal and aimed at general-purpose chat, this text-only variant is positioned specifically for harder reasoning and evaluation workloads, taking advantage of a 1,000,000-token combined context window that dwarfs the smaller model's 256K budget. The provider's documentation frames both models as uncensored systems that think before answering by default and stream their responses, with reasoning effort and structured outputs available as first-class controls.

In practice, the model is accessed through an OpenAI-compatible chat completions endpoint, so it can be dropped into existing agent frameworks with standard message-based calls, and it exposes tool and function calling alongside JSON-mode and JSON Schema structured outputs for verifiable pipelines. The combination of the cataloged API limit, generous output budget, reasoning control, and tool use makes it a fit for long-document analysis, multi-step agent workflows, and evaluation harnesses where the smaller multimodal default would feel constrained. It is less suitable when image or video inputs are required, since those capabilities are explicitly reserved for the standard Abliterated Model rather than this large text-focused variant.