Best Uncensored LLMs
Updated hourly · OpenRouter checked November 10, 2026

Uncensored LLMs on OpenRouter, and the ones it doesn't have

OpenRouter is the biggest model API, but uncensored models are a small corner of it: of its 469 models, one is explicitly uncensored and about ten are roleplay fine-tunes.

Short answerFor roleplay, OpenRouter has the community favourites (Hermes 4 405B, Magnum, Euryale). For uncensored versions of today's strongest open models (DeepSeek V4.1 Flash, GLM 5.3, Qwen, Gemma) you need another host. The best ranked of those is DeepSeek V4.1 Flash 1017 Uncensored.

Every uncensored or roleplay model on OpenRouter

ModelKindContext$ / 1M in · out
Dolphin Mistral 24B Venice Edition
cognitivecomputations/dolphin-mistral-24b-venice-edition
Uncensored128K$0.20 · $0.90
Hermes 4 405B
nousresearch/hermes-4-405b
Roleplay / general131K$1.00 · $3.00
Magnum v4 72B
anthracite-org/magnum-v4-72b
Roleplay33K$2.50 · $5.00
Euryale 70B v2.2
sao10k/l3.1-euryale-70b
Roleplay131K$0.85 · $0.85
Cydonia 24B v4.1
thedrummer/cydonia-24b-v4.1
Roleplay131K$0.30 · $0.50
Skyfall 36B v2
thedrummer/skyfall-36b-v2
Roleplay33K$0.55 · $0.80
UnslopNemo 12B
thedrummer/unslopnemo-12b
Roleplay1M$0.40 · $0.40
MythoMax 13B
gryphe/mythomax-l2-13b
Roleplay8K$0.08 · $0.11
ReMM SLERP 13B
undi95/remm-slerp-l2-13b
Roleplay6K$0.35 · $0.65
Lunaris 8B
sao10k/l3-lunaris-8b
Roleplay8K$0.04 · $0.05
Weaver
mancer/weaver
Roleplay8K$0.40 · $0.75

Prices are OpenRouter's list prices for these models. All of them are also served through OpenUncensored, which runs this site.

The best uncensored models OpenRouter doesn't list

Ranked by the OpenUncensored Score, with every host that serves each model.

RankModelScoreAgent tasksLoopsAnswers edgyHard-line refusals$ / 1M in+outWhere to run it
1DeepSeek V4.1 Flash 1017 Uncensored
Built on DeepSeek V4.1 Flash 1017 · deepseek-v4.1-flash-1017-uncensored
69.82/50%100%0%$1.65VeniceNanoGPTOpenUncensored
2GLM 5.3 Flash Uncensored
Built on GLM 5.3 Flash · glm-5.3-flash-uncensored
69.73/540%94%14%$1.00NanoGPTOpenUncensored
3Qwen 3.8 27B Uncensored Fast by Heabsy
Built on Qwen 3.8 27B · qwen3.8-27b-uncensored-fast
66.82/560%100%0%$4.00NanoGPTOpenUncensored
4DeepSeek V4.1 Flash 0910 Uncensored
Built on DeepSeek V4.1 Flash · deepseek-v4.1-flash-0910-uncensored
66.12/520%97%0%$1.65NanoGPTOpenUncensored
5DeepSeek V4.1 Flash Uncensored
Built on DeepSeek V4.1 Flash · deepseek-v4.1-flash-uncensored
64.31/50%100%0%$1.65VeniceNanoGPTOpenUncensored
6Abliterated Large Flash
Built on DeepSeek V4.1 Flash 1017 · abliterated-large-flash
64.01/520%92%0%$3.00Abliteration.aiVeniceNanoGPTOpenUncensored
7MiMo V2.6 Flash Uncensored
Built on MiMo V2.6 Flash · mimo-v2.6-flash-uncensored
63.52/50%83%93%$2.00NanoGPTOpenUncensored
8Apex Flash 1 Abliterated
Built on GLM 5.3 Flash · glm-5.3-flash-apex-abliterated
63.23/50%86%100%$0.65NanoGPTOpenUncensored
9GLM 5.3 Uncensored
Built on GLM 5.3 · glm-5.3-uncensored
63.02/520%92%0%$3.50NanoGPTOpenUncensored
10MiMo V2.6 Flash Abliterated
Built on MiMo V2.6 Flash · mimo-v2.6-flash-abliterated
62.62/520%83%93%$2.00NanoGPTOpenUncensored

Questions

Does OpenRouter have uncensored models?

A few. Of its 469 models on November 10, 2026, one is explicitly uncensored (Dolphin Mistral 24B Venice Edition) and about ten are roleplay fine-tunes. It does not list abliterated or heretic versions of DeepSeek, GLM, Qwen or Gemma.

Where can I run the abliterated models OpenRouter doesn't have?

On hosts such as Venice, NanoGPT and abliteration.ai, or through OpenUncensored, which routes across them on one OpenAI compatible key. Each model in the table below lists every host we know serves it.

See also: every host that serves uncensored models, and the best value uncensored APIs.