There are now 5 inference providers for abliterated DeepSeek

AllModels presents a routing service for abliterated, or uncensored, language models, highlighting DeepSeek V4.1 Flash Abliterated and GLM 5.3 Abliterated. DeepSeek is described as a fast reasoning model available through several independent inference providers; the page’s title points to five providers, though it does not name them. Multiple GLM deployments likewise share a public model ID. The central pitch is that developers can access these models through one API rather than integrating with each provider separately, using an OpenAI-compatible endpoint and familiar JavaScript or Python SDKs.

The router aims to keep requests available by automatically switching among eligible providers when one is overloaded or offline, without requiring code or model-ID changes. It sends image requests only to providers that support image input, while text requests can use the broader set of routes, and it tracks provider pricing to offer lower-cost options. The API supports both Chat Completions and Responses features, including streaming, tools, structured output, and reasoning controls. It also offers router-hosted web search, giving models a consistent search interface and URL citations regardless of their inference provider. The page frames DeepSeek and GLM as an initial selection within a growing catalogue, and invites users to browse more models or try the service.