Guides
Set up a coding tool, connect a provider's models, or learn how uncensored and abliterated variants differ. Every guide uses the exact model IDs and API paths RouterHub serves.
Coding tools
Providers
Uncensored models
- Uncensored vs abliteratedCompare uncensored and abliterated models, prompt-level and weight-level changes, exact DeepSeek, GLM and Qwen IDs, quality and cost.
- What is abliteration?Learn how abliteration changes model weights, how it differs from prompting or fine-tuning, and how to use abliterated API models.
- First uncensored API callConnect uncensored and abliterated models with an OpenAI-compatible SDK.
- Evaluating quality and costBuild repeatable model comparisons: refusal behavior, coding quality, tool use, reasoning settings, token costs and retained failures, with public GLM evidence.
- Hosted or local?Compare hosted uncensored APIs and local open-weight inference by hardware, privacy, deployment effort and total cost.
- DeepSeek uncensoredCompare DeepSeek model variants, exact API IDs, token rates, instruction behavior and compatible coding tools.
- GLM uncensoredCompare GLM model variants, exact API IDs, token rates, instruction behavior and compatible coding tools.
- Qwen uncensoredCompare Qwen model variants, exact API IDs, token rates, instruction behavior and compatible coding tools.
- Kimi uncensoredCompare Kimi model variants, exact API IDs, token rates, instruction behavior and compatible coding tools.