This uncensored LLM leaderboard ranks the 135 uncensored chat models you can call through the uncensored LLM API today. Two inputs, both public: the daily real-model check, which decides whether a model is on the board, and the live catalog price per million output tokens, which decides its place on it. Nothing on this page is typed by hand. Every name, price, context window and verdict is read at render, and the page rebuilds at least every five minutes.
125 of 135 models passed the check generated 2026-10-24 20:47 UTC. Prices are from the catalog of 2026-10-24.
Top of the board today
The lowest-priced model that answered the Oct 24 check as itself is GLM-5.3 Flash Uncensored at $0.3802 per million output tokens. 8 of the 125 models on the board are under one dollar per million output tokens.
| # | Model | Build | Output $/M | Context | Check |
|---|---|---|---|---|---|
| 1 | GLM-5.3 Flash Uncensored | Uncensored | $0.3802 | Steady since Oct 12 | |
| 2 | MythoMax L2 13B | Roleplay | $0.505 | 4K | Steady since Oct 23 |
| 3 | GLM-4.7 Flash Uncensored | Uncensored | $0.60 | Steady since Oct 19 | |
| 4 | Roleplay Uncensored | Roleplay | $0.60 | Steady since Oct 12 | |
| 5 | Cydonia 24B v4.3 | Roleplay | $0.75 | 131K | Steady since Oct 23 |
| 6 | Manta Flash 1.0 | Roleplay | $0.80 | 16K | Answered the check |
| 7 | Manta Mini 1.0 | Roleplay | $0.80 | 8K | Answered the check |
| 8 | GLM-5.3 Uncensored | Uncensored | $0.873 | Steady since Oct 12 | |
| 9 | Stheno 8B v3.2 | Roleplay | $1.005 | 16K | Steady since Oct 23 |
| 10 | GLM-5.3 Flash Abliterated | Abliterated | $1.10 | Steady since Oct 23 |
How this uncensored LLM leaderboard is ranked
Every published leaderboard for uncensored models ranks weights you download. The UGI Leaderboard by DontPlanToEnd on Hugging Face is the best of them, and this page does not try to replace it. It answers a different question: of the uncensored models you can call by API right now, which ones answered as themselves today, and what does each cost.
Input one: the daily real-model check
Once a day the gateway sends the same fixed prompt to every model in the catalog and records the answer. The prompt is public; the /proof endpoint prints it as the canary. Each model gets one of three verdicts. Steady: the answer matches the way that build has answered before, and the row shows since when. Changed: the answer moved on a given day. Not checked yet: the model is new to the catalog or missed the run.
The check proves two things. The model answering a request is the model named in it, not a swapped-in base model or a smaller build. And it was up when the check ran. It proves nothing about how uncensored the model is. There is no refusal count in it, no W/10 and no quality score. A heavily aligned stock model and a fully abliterated one both pass if they answer as themselves.
Input two: the live catalog
The catalog carries each model’s id, build tags, price per million input and output tokens, context window and wire formats. The price in the table is the price your key is billed at. Context windows come from the developer’s own page and show as blank when no page states one.
The ranking rule
- Models with a Steady verdict today are on the board. Changed and Not checked yet rows are listed separately below it.
- On the board, lower output price per million tokens ranks higher. Output is the side that dominates a chat or roleplay bill.
- Two models at the same output price keep alphabetical order. There is no tie-break on input price, context or build.
Price is the second key, not a quality score, for a plain reason: it is the only number on the row that is a fact rather than a judgment. A benchmark score depends on the prompts, the sampler and the judge, and every published uncensored benchmark rates weights under its own settings, not the served build. The price is what the key is billed, to the fraction of a cent, and it changes only when the catalog does. Ranking on it also makes the board useful for the most common question behind the query, which is where to send a roleplay or writing workload without paying flagship rates for a model that was never going to refuse anyway.
What the board will not do: average the two inputs into one score, accept self-reported numbers from a model author, or take payment for a position. There are no sponsored rows. A model moves up by passing the check and costing less, and down by failing it or costing more.
That is the whole method. It is deliberately narrow so that every row is a fact you can re-derive. The two endpoints need no key, and the join is a few lines of shell.
# 1. Which models answered today's check as the build they are named forcurl -s https://api.uncensoredllm.ai/v1/proof \ | jq -r '.data[] | select(.verdict == "stable") | .model' > stable.txt # 2. Their prices and context windows, lowest output price firstcurl -s https://api.uncensoredllm.ai/v1/catalog \ | jq -r --rawfile ok stable.txt '($ok | split("\n")) as $s | .data[] | select((.tags // []) | index("uncensored")) | select(.id as $i | $s | index($i)) | [.id, .pricing.output, .context_window] | @tsv' \ | sort -t$'\t' -k2,2nThe complementary measure is UGI. The UGI score grades how much a model knows about, and will discuss, topics that aligned models decline; W/10 is a willingness score out of ten, kept separate from the knowledge score so a model that knows little but refuses nothing does not look clever. Use UGI to decide how uncensored a model is. Use this board to find out whether you can call it today and what it costs.
The board
125 models answered the Oct 24 check as the build they are named for. Sorted by output price; click any header to re-sort. The model column links to each model’s page with its first-request example. Prices are per million tokens as the price list shows them.
| Wire formats | |||||
|---|---|---|---|---|---|
| GLM-5.3 Flash Uncensored | Uncensored | $0.14 | $0.3802 | ||
| MythoMax L2 13B | Roleplay | $0.505 | $0.505 | 4K | |
| GLM-4.7 Flash Uncensored | Uncensored | $0.12 | $0.60 | ||
| Roleplay Uncensored | Roleplay | $0.30 | $0.60 | ||
| Cydonia 24B v4.3 | Roleplay | $0.605 | $0.75 | 131K | |
| Manta Flash 1.0 | Roleplay | $0.10 | $0.80 | 16K | |
| Manta Mini 1.0 | Roleplay | $0.10 | $0.80 | 8K | |
| GLM-5.3 Uncensored | Uncensored | $0.30 | $0.873 | ||
| Stheno 8B v3.2 | Roleplay | $1.005 | $1.005 | 16K | |
| GLM-5.3 Flash Abliterated | Abliterated | $1.10 | $1.10 | ||
| Gemma 4 12B Starry Ocean | Roleplay | $0.25 | $1.25 | 131K | |
| Gemma 4 12B StationKeeper | Roleplay | $0.25 | $1.25 | 131K | |
| Amoral Gemma3 27B v2 | Uncensored | $1.50 | $1.50 | 33K | |
| Gemma 4 31B Split-Untied | Roleplay | $0.50 | $1.50 | 262K | |
| Veiled Calla 12B | Roleplay | $1.50 | $1.50 | 33K | |
| Gemma 4 31B Agares v1 | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Animus V14.1 | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B AssGuard | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Cognitive Unshackled | Abliterated | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Dark Gemistry | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Garnet V2 | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Gembrain Heretic | Abliterated | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Gembrain X Core | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Isometry RP | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B K1 v5 | Uncensored | $1.53 | $1.53 | 262K | |
| Gemma 4 31B Queen | Roleplay | $1.53 | $1.53 | 262K | |
| Gemma 4 31B SDFT Heretic RP | Abliterated | $1.53 | $1.53 | 262K | |
| Gemma 4 31B StyleTune | Roleplay | $1.53 | $1.53 | 262K | |
| Qwen3.5 27B Blossom V6.4 Derestricted | Abliterated | $1.53 | $1.53 | 262K | |
| Qwen3.5 27B BlueStar v3 Derestricted | Abliterated | $1.53 | $1.53 | 262K | |
| Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted | Abliterated | $1.53 | $1.53 | 262K | |
| Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted Lite | Abliterated | $1.53 | $1.53 | 262K | |
| Qwen3.5 27B Queen Derestricted | Abliterated | $1.53 | $1.53 | 262K | |
| Anubis 70B v1.1 | Roleplay | $1.55 | $1.55 | 32K | |
| DeepSeek V4.1 Flash Abliterated | Abliterated | $1.75 | $1.75 | ||
| Euryale 70B v2.2 | Roleplay | $1.53 | $1.785 | 20K | |
| GPT OSS 120B Uncensored | Uncensored | $0.365 | $1.82 | 131K | |
| Gemma 4 26B-A4B ChimeraX | Roleplay | $0.60 | $1.90 | 262K | |
| Gemma 4 26B-A4B ChimeraX Heretic | Abliterated | $0.605 | $1.90 | 262K | |
| Gemma 4 26B-A4B DarkSoul | Roleplay | $0.605 | $1.90 | 262K | |
| Gemma 4 26B-A4B Luminous Mirror | Roleplay | $0.60 | $1.90 | 262K | |
| Gemma 4 26B-A4B MeroMero | Roleplay | $0.60 | $1.90 | 262K | |
| Gemma 4 26B-A4B Moonlight Dusk | Roleplay | $0.60 | $1.90 | 262K | |
| Gemma 4 26B-A4B Musica | Roleplay | $0.60 | $1.90 | 262K | |
| Gemma 4 26B-A4B Opus Distill | Roleplay | $0.60 | $1.90 | 262K | |
| Gemma 4 26B-A4B Shadow Siren | Roleplay | $0.60 | $1.90 | 262K | |
| DeepSeek V4 Pro Uncensored | Uncensored | $1.25 | $2.00 | ||
| Hermes 3 Llama 3.1 70B | Uncensored | $2.04 | $2.04 | 66K | |
| NeuralDaredevil 8B Abliterated | Abliterated | $2.20 | $2.20 | 8K | |
| Artemis v1.1 | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B DarkIdol | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B Fabled | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B Gembrain | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B Gemsicle | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B Isometry | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B MeroMero v2 | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B Novelist | Roleplay | $0.50 | $2.25 | 262K | |
| Gemma 4 31B Scarlet | Roleplay | $0.50 | $2.25 | 262K | |
| Cu-Mai R1 70B | Roleplay | $2.465 | $2.465 | 33K | |
| Euryale 70B v2.3 | Roleplay | $2.465 | $2.465 | 20K | |
| Hanami 70B x1 | Roleplay | $2.465 | $2.465 | 33K | |
| Mag Mell R1 12B | Roleplay | $2.465 | $2.465 | 16K | |
| Magnum Diamond 24B | Roleplay | $2.465 | $2.465 | 33K | |
| NemoMix Unleashed 12B | Roleplay | $2.465 | $2.465 | 33K | |
| Starcannon Unleashed 12B v1.0 | Roleplay | $2.465 | $2.465 | 16K | |
| UnslopNemo 12B v4.1 | Roleplay | $2.465 | $2.465 | 8K | |
| Cydonia 24B v4.1 | Roleplay | $1.50 | $2.50 | 131K | |
| GLM-5.3 Flash Apex Abliterated | Abliterated | $0.75 | $2.50 | 1049K | |
| GLM-5.3 Flash Cybersecurity | Security | $0.75 | $2.50 | 1049K | |
| Manta Pro 1.0 | Roleplay | $0.30 | $2.50 | 66K | |
| The Omega Directive 24B v2.0 | Roleplay | $2.50 | $2.50 | 33K | |
| Uncensored AI Novelist | House build | $1.53 | $3.00 | 262K | |
| Qwen3.8 27B Antislop | Roleplay | $0.50 | $3.00 | 262K | |
| Qwen3.8 27B DarkIdol | Roleplay | $0.50 | $3.00 | 262K | |
| Qwen3.8 27B Fimi | Roleplay | $0.50 | $3.00 | 262K | |
| Qwen3.8 27B Kimiko | Roleplay | $0.50 | $3.00 | 262K | |
| Qwen3.8 27B Kiwen | Roleplay | $0.50 | $3.00 | 262K | |
| Qwen3.8 27B Novelist | Roleplay | $0.50 | $3.00 | 262K | |
| Qwen3.8 27B Synthia | Roleplay | $0.50 | $3.00 | 262K | |
| DeepSeek R1 Distill Llama 70B Abliterated | Abliterated | $3.50 | $3.50 | 16K | |
| Gemma 4 26B-A4B Uncensored | Uncensored | $0.75 | $3.50 | 33K | |
| Llama 3.3 70B Abliterated | Abliterated | $3.50 | $3.50 | 33K | |
| Jailbreak-Free Assistant | House build | $1.00 | $4.00 | 1049K | |
| Skyfall 36B v2 | Roleplay | $2.75 | $4.00 | 33K | |
| Uncensored Coding Agent | House build | $1.00 | $4.00 | 1049K | |
| DeepSeek V4 Flash 0731 Uncensored | Uncensored | $1.455 | $4.365 | 262K | |
| GLM-5.3 Abliterated | Abliterated | $3.75 | $4.50 | ||
| ReMM SLERP L2 13B | Roleplay | $2.545 | $4.725 | 6K | |
| QwQ 32B Uncensored | Uncensored | $3.30 | $5.00 | 24K | |
| Hermes 4 405B | Uncensored | $1.50 | $6.00 | 128K | |
| DeepSeek R1 Distill Qwen 32B Abliterated | Abliterated | $7.00 | $7.00 | 16K | |
| MiMo V2.6 Flash Abliterated | Abliterated | $2.50 | $7.50 | 1049K | |
| MiMo V2.6 Flash Uncensored | Uncensored | $2.50 | $7.50 | 1049K | |
| Qwen3.8 27B Fable | Roleplay | $1.25 | $7.50 | 524K | |
| Qwen3.8 27B Hemingway | Roleplay | $1.25 | $7.50 | 262K | |
| Qwen3.8 27B Queen | Roleplay | $1.25 | $7.50 | 524K | |
| Aion RP Llama 3.1 8B | Roleplay | $4.00 | $8.00 | 33K | |
| Qwen3.8 27B Obliterated | Abliterated | $1.25 | $8.75 | 524K | |
| Dolphin Mistral 24B Venice Edition | Uncensored | $2.00 | $9.00 | 33K | |
| Uncensored AI Roleplay | Roleplay | $2.00 | $9.00 | 128K | |
| Chat Uncensored | Uncensored | $5.00 | $10.00 | ||
| Hy3 Uncensored | Uncensored | $2.50 | $10.00 | 262K | |
| MiMo-V2.5 Uncensored | Uncensored | $2.50 | $10.00 | 262K | |
| MiniMax M3 Uncensored | Uncensored | $2.50 | $10.00 | 262K | |
| GLM 5.2 Uncensored | Uncensored | $3.65 | $11.70 | 262K | |
| DansPE 1.3 24B | Roleplay | $3.635 | $14.54 | 33K | |
| Magnum v4 72B | Roleplay | $10.03 | $14.96 | 16K | |
| Qwen3.8 27B Uncensored | Uncensored | $5.00 | $15.00 | 33K | |
| GLM 4.7 Uncensored | Uncensored | $5.09 | $18.175 | 131K | |
| Kimi K2.6 Uncensored | Uncensored | $3.90 | $18.35 | 262K | |
| No Filter AI | House build | $19.50 | $19.50 | 197K | |
| NSFW AI Chat | House build | $19.50 | $19.50 | 131K | |
| Uncensored AI Assistant | House build | $19.50 | $19.50 | 197K | |
| Uncensored AI Companion | House build | $19.50 | $19.50 | 131K | |
| Uncensored AI Writer | House build | $19.50 | $19.50 | 131K | |
| Uncensored Chatbot | House build | $19.50 | $19.50 | 197K | |
| Unfiltered AI | House build | $19.50 | $19.50 | 197K | |
| GLM-5.3 Flash Code Uncensored | Uncensored | $20.00 | $20.00 | ||
| CyberHeabsy | Security | $7.50 | $22.50 | 1049K | |
| Gemma 31B Heretic Uncensored | Abliterated | $5.00 | $27.50 | ||
| Code Pro Uncensored | Uncensored | $15.00 | $30.00 | ||
| Uncensored AI Chat | House build | $39.00 | $39.00 | 197K | |
| Uncensored Code Assistant | House build | $39.00 | $39.00 | 131K | |
| Unrestricted AI | House build | $39.00 | $39.00 | 131K | |
| Kimi K3 Uncensored | Uncensored | $19.15 | $95.60 | 262K |
Not ranked today
10 models are off the board: 0 answered differently from before and 10 were not checked in this run. They are still callable; a missing verdict is not a verdict.
| Model | Build | Output $/M | Check |
|---|---|---|---|
| Magidonia 24B v4.3 | Roleplay | $0.605 | Not checked yet |
| Red Team (Flash) | Security | $3.00 | Not checked yet |
| Llama 3 70B Abliterated v3.5 | Abliterated | $3.50 | Not checked yet |
| Wayfarer Large 70B | Roleplay | $3.505 | Not checked yet |
| Red Team (Pro) | Security | $6.00 | Not checked yet |
| GLM-4.6 Derestricted v5 | Abliterated | $7.50 | Not checked yet |
| Laguna S 2.1 Uncensored | Uncensored | $10.00 | Not checked yet |
| MiMo-V2.5-Pro Uncensored | Uncensored | $10.00 | Not checked yet |
| Nemotron 3 Ultra 550B Uncensored | Uncensored | $16.50 | Not checked yet |
| Inkling Uncensored | Uncensored | $30.35 | Not checked yet |
The models page carries the same rows without the check column, in order of paid usage rather than price.
Notable rows
- Lowest output price on the board: GLM-5.3 Flash Uncensored (uncensored build) at $0.3802 per million output tokens and $0.14 input.
- Largest context window on the board: GLM-5.3 Flash Apex Abliterated at 1,048,576 tokens, $2.50 per million output tokens.
- Longest unbroken run: GLM-5.3 Flash Uncensored, steady since oct 12.
- Builds on the board: 20 abliterated, 25 uncensored, 65 roleplay, 2 security, 13 house build.
Build labels matter when you read the price column. An abliterated model had its refusal direction removed from the weights after training, which is why a 27B abliterated build and its stock twin share a price. An uncensored fine-tune was retrained without refusals. A roleplay build was tuned on long-form character chat and is usually a smaller model at a lower price. The difference between the first two is the subject of abliterated vs uncensored; Maxime Labonne’s abliteration write-up on Hugging Face is the reference for the technique itself.
Two models at the same price are not the same model. The board tells you the price and that the model is itself today. Whether it writes well for your use is a separate question, and the best uncensored LLM by API in 2026 post ranks by use rather than by price.
Using the board with UGI
The two leaderboards answer different questions, so use them in sequence. Pick candidates on UGI by score and parameter count. Then find the same family on this board. The row gives you the id, the price and today’s verdict. The board carries community fine-tunes of Gemma 4, Qwen 3.5 and Llama 3 bases under the names their authors gave them, so a name you found on UGI is searchable here as is.
Calling a model from the board takes one request. The id in the model column is the id you send. The request shape is the standard chat completions body, so any OpenAI-compatible client works without changes beyond the base URL and the key.
curl https://api.uncensoredllm.ai/v1/chat/completions \ -H "Authorization: Bearer sk-ullm-live-…" \ -H "Content-Type: application/json" \ -d '{"model": "glm-5.3-flash-uncensored", "messages": [{"role": "user", "content": "Hello"}]}'Base URL https://api.uncensoredllm.ai/v1. The key comes from the console and the request settles per token at the price in the row. SillyTavern, Janitor AI, OpenCode and Cline settings are on the docs page.
A worked example from the same catalog read. 1,000 turns of 2,000 input tokens and 300 output tokens each, which is a long evening of roleplay with a full character card in context, costs $0.39 on GLM-5.3 Flash Uncensored and $66.98 on Kimi K3 Uncensored, the highest-priced model on today’s board. Same board, same check, a 170.0x spread in price. The balance is prepaid and settles per request at these rates, and the console shows the receipt for each call.
Re-check the board before a long job. The verdict column changes once a day, the price column whenever the catalog does, and the page rebuilds within five minutes of either. The date line at the top of the page is the date to quote.
Questions
Yes. The UGI Leaderboard on Hugging Face, by DontPlanToEnd, scores open-weight models on uncensored general intelligence and willingness. This page is the complementary board for models you can call by API: it ranks them by the daily real-model check and then by output price. See the daily check.
UGI stands for Uncensored General Intelligence. The UGI Leaderboard is a Hugging Face Space that ranks language models by a UGI score, which grades knowledge of and willingness to discuss topics that aligned models decline, and by W/10, a willingness score out of ten. It tests weights you run yourself.
Open the Space, sort by UGI or W/10, and filter by parameter count to fit your hardware or budget. To call a listed model without downloading it, find the same name on the models page here and send its id in a chat completions request. See the models page.
W/10 is the willingness score, out of ten: how readily a model answers sensitive questions. It is kept separate from the UGI knowledge score, so a model that refuses nothing but knows little does not rank as if it were capable.
It depends on the job. On the board of 2026-10-24, the lowest-priced model that passed the real-model check is GLM-5.3 Flash Uncensored at $0.3802 per million output tokens, and GLM-5.3 Flash Apex Abliterated has the largest context window at 1,048,576 tokens. See the best uncensored LLM by API in 2026.
Hugging Face hosts arena-style text-to-image leaderboards as Spaces. They score image quality by human preference, not whether a model runs with its safety checker off. Uncensored image and video models you can call by API are listed with per-image and per-second prices on the media page. See the media page.