From the Uplink Switch LLC family: this post first appeared on UncensoredLLM at uncensoredllm.ai/blog/uncensored-llm-leaderboard, where the figures are read live from the catalog.
The best uncensored LLMs you can call by API, ranked by today’s real-model check, then by output price per million tokens. UGI scores beside the rows UGI has tested.
Published October 25, 2026 · prices as of October 27, 2026
This leaderboard ranks the best uncensored LLMs you can call by API on one key: 136 uncensored LLMs today. A model gets on the board by passing the daily real-model check. Its place on the board is its output price per million tokens. Where the UGI Leaderboard has tested the same model, its UGI score sits beside the row.
127 of 136 models passed the check run at 2026-10-27 00:20 UTC. Prices are from the catalog of October 27, 2026. Every figure on this page is read at render, and the page rebuilds at least every five minutes.
Uncensored LLM leaderboard, October 2026
GLM-5.3 Flash Uncensored leads the October 2026 board at $0.3802 per million output tokens. 14 of the 127 models on the board cost 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 | $0.404 | 4K | Steady since Oct 20 |
| 3 | Magidonia 24B v4.3 | Roleplay | $0.4235 | 33K | Steady since Oct 26 |
| 4 | Manta Flash 1.0 | Roleplay | $0.56 | 16K | Answered the check |
| 5 | Manta Mini 1.0 | Roleplay | $0.56 | 8K | Answered the check |
| 6 | Cydonia 24B v4.3 | Roleplay | $0.60 | 131K | Steady since Oct 20 |
| 7 | GLM-4.7 Flash Uncensored | Uncensored | $0.60 | Steady since Oct 22 | |
| 8 | Roleplay Uncensored | Roleplay | $0.60 | Steady since Oct 12 | |
| 9 | Stheno 8B v3.2 | Roleplay | $0.7035 | 16K | Steady since Oct 21 |
| 10 | GPT OSS 120B Uncensored | Uncensored | $0.728 | 131K | Steady since Oct 21 |
Top 10 of 127. Check run 2026-10-27 00:20 UTC, prices from the catalog of 2026-10-27. The month in this heading changes on the first of each month. The URL stays the same.
Picks by job
The board ranks on price. These picks read the same rows for a job instead. Each one is the first row of its kind on today’s board, or the UGI figure named beside it.
| If you want | Pick | Why, from today’s board |
|---|---|---|
| The lowest bill per reply | GLM-5.3 Flash Uncensored | Lowest output price on the board, $0.3802 per million tokens |
| No refusal direction in the weights | GLM-5.3 Flash Abliterated | Lowest-priced abliterated build that passed, $0.88 per million output tokens |
| Character chat from a card client | MythoMax L2 13B | Lowest-priced roleplay build that passed, $0.404 per million output tokens |
| Security research | GLM-5.3 Flash Cybersecurity | Lowest-priced security build that passed, $1.75 per million output tokens |
| A whole repository or book in one request | GLM-5.3 Flash Cybersecurity | Largest context window on the board, 1,048,576 tokens |
| The id with the longest steady run | GLM-5.3 Flash Uncensored | Steady since Oct 12, the longest unbroken run on the board |
| The highest UGI score on the board | Hermes 4 405B | UGI 54.07, W/10 4.8 on the UGI Leaderboard (tested September 21, 2025) |
| The most willing on UGI | Anubis 70B v1.1 | W/10 8.5, the highest willingness of the UGI rows on the board (tested September 10, 2025) |
The build label decides what the price buys. An abliterated build had its refusal direction removed from the weights after training. An uncensored build was fine-tuned to answer. A roleplay tune was trained on long-form character chat and is often a smaller model at a lower price. Abliterated vs uncensored covers the difference between the first two.
Best abliterated LLM on the board
21 of the 23 abliterated builds on the catalog passed the Oct 27 check. They rank the same way as the board: passed first, then output price. Heretic, derestricted and unshackled builds count here, because the catalog tags every build whose refusal was removed from the weights as abliterated.
| # | Model | Output $/M | Context | Check |
|---|---|---|---|---|
| 1 | GLM-5.3 Flash Abliterated | $0.88 | Steady since Oct 22 | |
| 2 | Gemma 4 31B Cognitive Unshackled | $1.071 | 262K | Steady since Oct 20 |
| 3 | Qwen3.5 27B Blossom V6.4 Derestricted | $1.071 | 262K | Steady since Oct 21 |
| 4 | Qwen3.5 27B BlueStar v3 Derestricted | $1.071 | 262K | Steady since Oct 21 |
| 5 | Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted | $1.071 | 262K | Steady since Oct 21 |
| 6 | Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted Lite | $1.071 | 262K | Steady since Oct 21 |
| 7 | Qwen3.5 27B Queen Derestricted | $1.071 | 262K | Steady since Oct 21 |
| 8 | Gemma 4 31B Gembrain Heretic | $1.224 | 262K | Steady since Oct 20 |
The rest of the abliterated builds are in the full board below, with the same columns.
The board next to UGI
The UGI Leaderboard by DontPlanToEnd on Hugging Face measures how uncensored a model is. Its UGI score combines knowledge of sensitive topics with W/10, short for willingness out of ten: how far a model can be pushed before it refuses or drifts from instructions. Its test questions are kept private, and it scores open-weight and proprietary models.
UGI does not say whether a build is callable by API today or what it costs per token. That is this board’s job. 10 models on the catalog have a UGI row under the same author and model name, highest UGI first:
| Model | UGI row | UGI | W/10 | Check today | Output $/M |
|---|---|---|---|---|---|
| Hermes 4 405B | NousResearch/Hermes-4-405B |
54.07 | 4.8 | Steady since Oct 20 | $4.80 |
| Anubis 70B v1.1 | TheDrummer/Anubis-70B-v1.1 |
50.79 | 8.5 | Steady since Oct 20 | $1.085 |
| Cydonia 24B v4.3 | TheDrummer/Cydonia-24B-v4.3 |
44.52 | 7.0 | Steady since Oct 20 | $0.60 |
| Dolphin Mistral 24B Venice Edition | dphn/Dolphin-Mistral-24B-Venice-Edition |
40.79 | 7.8 | Steady since Oct 20 | $2.25 |
| Wayfarer Large 70B | LatitudeGames/Wayfarer-Large-70B-Llama-3.3 |
37.06 | 5.0 | Not checked yet | $2.4535 |
| UnslopNemo 12B v4.1 | TheDrummer/UnslopNemo-12B-v4.1 |
34.17 | 6.2 | Steady since Oct 20 | $1.972 |
| Skyfall 36B v2 | TheDrummer/Skyfall-36B-v2 |
33.90 | 6.0 | Steady since Oct 20 | $2.80 |
| NeuralDaredevil 8B Abliterated | mlabonne/NeuralDaredevil-8B-abliterated |
33.20 | 6.5 | Steady since Oct 24 | $1.76 |
| Mag Mell R1 12B | inflatebot/MN-12B-Mag-Mell-R1 |
31.14 | 6.0 | Steady since Oct 23 | $1.7255 |
| Stheno 8B v3.2 | Sao10K/L3-8B-Stheno-v3.2 |
20.56 | 3.8 | Steady since Oct 21 | $0.7035 |
UGI and W/10 from the UGI Leaderboard’s data file (last changed October 17, 2026), read October 25, 2026. UGI tested these weights between September 8, 2025 and December 16, 2025. Its scores describe the weights as UGI ran them; the check column describes the build this key calls today.
How this uncensored LLM leaderboard is ranked
Two inputs, both public, and no score that blends them into one.
Input one: the daily real-model check
Once a day the gateway sends one fixed prompt to every model on the catalog and records the answer. Each model gets one of three verdicts. Steady: the answer matches how that build has answered before, with the date it has held since. Changed: the answer moved on the date shown. Not checked yet: the model is new to the catalog or missed the run.
The check shows that the model answering is the model named in the request, and that it was up when the check ran. It does not measure how uncensored the model is. A stock model and an abliterated one both pass if they answer as themselves. UGI covers that question.
Input two: the live catalog
The catalog carries each model’s id, build tags, price per million input and output tokens and context window. The price in the table is the price the key is billed. A blank context means the catalog has no figure for that model.
The ranking rule
- Models with a Steady verdict today are on the board. Changed and Not checked yet rows are listed below it.
- On the board, the 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 because it is the one number on the row that is a fact. A benchmark score depends on the prompts, the sampler and the judge. The price is what the key is billed, to the fraction of a cent, and it changes only when the catalog does.
No sponsored rows, no numbers reported by a model’s author and no paid positions. A model moves up by passing the check and costing less, and down by failing it or costing more.
Both data files need no key. These two commands rebuild the board:
# 1. Models that answered today's check as the build they are named for
curl -s https://uncensoredllm.ai/proof.json \
| jq -r '.data[] | select(.verdict == "stable") | .model' > stable.txt
# 2. Their output prices and context windows, lowest output price first
curl -s https://uncensoredllm.ai/pricing-index.json \
| jq -r --rawfile ok stable.txt '($ok | split("\n")) as $s
| .data[]
| select(.id | IN($s[]))
| [.id, .output, .context_window] | @tsv' \
| sort -t$'\t' -k2,2n
The full board
127 models answered the Oct 27 check as the build they are named for. The table sorts by output price; click any header to re-sort. Each name links to the model’s page and its first request. Prices are per million tokens, as on the price list.
| Model | Build | Input $/M | Output $/M | Context | Wire formats |
|---|---|---|---|---|---|
| GLM-5.3 Flash Uncensored glm-5.3-flash-uncensored | Uncensored | $0.1086 | $0.3802 | openai | |
| MythoMax L2 13B mythomax-l2-13b | Roleplay | $0.404 | $0.404 | 4K | openai |
| Magidonia 24B v4.3 magidonia-24b-v4.3 | Roleplay | $0.3535 | $0.4235 | 33K | openai |
| Manta Flash 1.0 manta-flash-1.0 | Roleplay | $0.07 | $0.56 | 16K | openai |
| Manta Mini 1.0 manta-mini-1.0 | Roleplay | $0.07 | $0.56 | 8K | openai |
| Cydonia 24B v4.3 cydonia-24b-v4.3 | Roleplay | $0.484 | $0.60 | 131K | openai |
| GLM-4.7 Flash Uncensored glm-4.7-flash-uncensored | Uncensored | $0.12 | $0.60 | openai | |
| Roleplay Uncensored roleplay-uncensored | Roleplay | $0.30 | $0.60 | openai | |
| Stheno 8B v3.2 stheno-8b-v3.2 | Roleplay | $0.7035 | $0.7035 | 16K | openai |
| GPT OSS 120B Uncensored gpt-oss-120b-uncensored | Uncensored | $0.1095 | $0.728 | 131K | openai |
| GLM-5.3 Uncensored glm-5.3-uncensored | Uncensored | $0.282 | $0.846 | openai | |
| Gemma 4 12B Starry Ocean gemma-4-12b-starry-ocean | Roleplay | $0.175 | $0.875 | 131K | openai |
| Gemma 4 12B StationKeeper gemma-4-12b-station-keeper | Roleplay | $0.175 | $0.875 | 131K | openai |
| GLM-5.3 Flash Abliterated glm-5.3-flash-abliterated | Abliterated | $0.88 | $0.88 | openai | |
| Amoral Gemma3 27B v2 amoral-gemma3-27b-v2 | Uncensored | $1.05 | $1.05 | 33K | openai |
| Gemma 4 31B Split-Untied gemma-4-31b-split-untied | Roleplay | $0.35 | $1.05 | 262K | openai |
| Veiled Calla 12B veiled-calla-12b | Roleplay | $1.05 | $1.05 | 33K | openai |
| Gemma 4 31B Agares v1 gemma-4-31b-agares-v1 | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B Animus V14.1 gemma-4-31b-animus-v14.1 | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B AssGuard gemma-4-31b-assguard | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B Cognitive Unshackled gemma-4-31b-cognitive-unshackled | Abliterated | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B Garnet V2 gemma-4-31b-garnet-v2 | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B Gembrain X Core gemma-4-31b-gembrain-x-core | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B Isometry RP gemma-4-31b-isometry-rp | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B K1 v5 gemma-4-31b-k1-v5 | Uncensored | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B Queen gemma-4-31b-queen | Roleplay | $1.071 | $1.071 | 262K | openai |
| Gemma 4 31B StyleTune gemma-4-31b-styletune | Roleplay | $1.071 | $1.071 | 262K | openai |
| Qwen3.5 27B Blossom V6.4 Derestricted qwen3.5-27b-blossom-v6.4-derestricted | Abliterated | $1.071 | $1.071 | 262K | openai |
| Qwen3.5 27B BlueStar v3 Derestricted qwen3.5-27b-bluestar-v3-derestricted | Abliterated | $1.071 | $1.071 | 262K | openai |
| Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted qwen3.5-27b-claude-4.6-opus-distilled-derestricted | Abliterated | $1.071 | $1.071 | 262K | openai |
| Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted Lite qwen3.5-27b-claude-4.6-opus-distilled-derestricted-lite | Abliterated | $1.071 | $1.071 | 262K | openai |
| Qwen3.5 27B Queen Derestricted qwen3.5-27b-queen-derestricted | Abliterated | $1.071 | $1.071 | 262K | openai |
| Anubis 70B v1.1 anubis-70b-v1.1 | Roleplay | $1.085 | $1.085 | 32K | openai |
| Gemma 4 31B Dark Gemistry gemma-4-31b-dark-gemistry | Roleplay | $1.224 | $1.224 | 262K | openai |
| Gemma 4 31B Gembrain Heretic gemma-4-31b-gembrain-heretic | Abliterated | $1.224 | $1.224 | 262K | openai |
| Gemma 4 31B SDFT Heretic RP gemma-4-31b-sdft-heretic-rp | Abliterated | $1.224 | $1.224 | 262K | openai |
| Gemma 4 26B-A4B Uncensored gemma-4-26b-a4b-uncensored | Uncensored | $0.27 | $1.26 | 33K | openai |
| Gemma 4 26B-A4B ChimeraX gemma-4-26b-a4b-chimerax | Roleplay | $0.42 | $1.33 | 262K | openai |
| Gemma 4 26B-A4B Luminous Mirror gemma-4-26b-a4b-luminous-mirror | Roleplay | $0.42 | $1.33 | 262K | openai |
| Gemma 4 26B-A4B MeroMero gemma-4-26b-a4b-meromero | Roleplay | $0.42 | $1.33 | 262K | openai |
| Gemma 4 26B-A4B Moonlight Dusk gemma-4-26b-a4b-moonlight-dusk | Roleplay | $0.42 | $1.33 | 262K | openai |
| Gemma 4 26B-A4B Musica gemma-4-26b-a4b-musica | Roleplay | $0.42 | $1.33 | 262K | openai |
| Gemma 4 26B-A4B Opus Distill gemma-4-26b-a4b-opus-distill | Roleplay | $0.42 | $1.33 | 262K | openai |
| Gemma 4 26B-A4B Shadow Siren gemma-4-26b-a4b-shadow-siren | Roleplay | $0.42 | $1.33 | 262K | openai |
| Euryale 70B v2.2 euryale-70b-v2.2 | Roleplay | $1.224 | $1.428 | 20K | openai |
| Gemma 4 26B-A4B ChimeraX Heretic gemma-4-26b-a4b-chimerax-heretic | Abliterated | $0.484 | $1.52 | 262K | openai |
| Gemma 4 26B-A4B DarkSoul gemma-4-26b-a4b-darksoul | Roleplay | $0.484 | $1.52 | 262K | openai |
| Artemis v1.1 artemis-v1.1 | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B Fabled gemma-4-31b-fabled | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B Gembrain gemma-4-31b-gembrain | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B Gemsicle gemma-4-31b-gemsicle | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B Isometry gemma-4-31b-isometry | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B MeroMero v2 gemma-4-31b-meromero-v2 | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B Novelist gemma-4-31b-novelist | Roleplay | $0.35 | $1.575 | 262K | openai |
| Gemma 4 31B Scarlet gemma-4-31b-scarlet | Roleplay | $0.35 | $1.575 | 262K | openai |
| Hermes 3 Llama 3.1 70B hermes-3-llama-3.1-70b | Uncensored | $1.632 | $1.632 | 66K | openai |
| DeepSeek V4.1 Flash Abliterated deepseek-v4.1-flash-abliterated | Abliterated | $0.8488 | $1.6975 | openai | |
| Cu-Mai R1 70B cu-mai-r1-70b | Roleplay | $1.7255 | $1.7255 | 33K | openai |
| Hanami 70B x1 hanami-70b-x1 | Roleplay | $1.7255 | $1.7255 | 33K | openai |
| Mag Mell R1 12B mag-mell-12b-r1 | Roleplay | $1.7255 | $1.7255 | 16K | openai |
| NemoMix Unleashed 12B nemomix-unleashed-12b | Roleplay | $1.7255 | $1.7255 | 33K | openai |
| Starcannon Unleashed 12B v1.0 starcannon-unleashed-12b | Roleplay | $1.7255 | $1.7255 | 16K | openai |
| DeepSeek V4 Flash 0731 Uncensored deepseek-v4-flash-0731-uncensored | Uncensored | $0.4365 | $1.746 | 262K | openai |
| GLM-5.3 Flash Cybersecurity glm-5.3-flash-cybersecurity | Security | $0.525 | $1.75 | 1049K | openai |
| Manta Pro 1.0 manta-pro-1.0 | Roleplay | $0.21 | $1.75 | 66K | openai |
| NeuralDaredevil 8B Abliterated neuraldaredevil-8b-abliterated | Abliterated | $1.76 | $1.76 | 8K | openai |
| Gemma 4 31B DarkIdol gemma-4-31b-darkidol | Roleplay | $0.40 | $1.80 | 262K | openai |
| Euryale 70B v2.3 euryale-70b-v2.3 | Roleplay | $1.972 | $1.972 | 20K | openai |
| Magnum Diamond 24B magnum-diamond-24b | Roleplay | $1.972 | $1.972 | 33K | openai |
| UnslopNemo 12B v4.1 unslopnemo-12b-v4.1 | Roleplay | $1.972 | $1.972 | 8K | openai |
| Cydonia 24B v4.1 cydonia-24b-v4.1 | Roleplay | $1.20 | $2.00 | 131K | openai |
| DeepSeek V4 Pro Uncensored deepseek-v4-pro-uncensored | Uncensored | $1.00 | $2.00 | openai | |
| GLM-5.3 Flash Apex Abliterated glm-5.3-flash-apex-abliterated | Abliterated | $0.60 | $2.00 | 1049K | openai |
| The Omega Directive 24B v2.0 omega-directive-24b-v2 | Roleplay | $2.00 | $2.00 | 33K | openai |
| Qwen3.8 27B Antislop qwen3.8-27b-antislop | Roleplay | $0.35 | $2.10 | 262K | openai |
| Qwen3.8 27B Fimi qwen3.8-27b-fimi | Roleplay | $0.35 | $2.10 | 262K | openai |
| Qwen3.8 27B Kimiko qwen3.8-27b-kimiko | Roleplay | $0.35 | $2.10 | 262K | openai |
| Qwen3.8 27B Kiwen qwen3.8-27b-kiwen | Roleplay | $0.35 | $2.10 | 262K | openai |
| Qwen3.8 27B Novelist qwen3.8-27b-novelist | Roleplay | $0.35 | $2.10 | 262K | openai |
| Qwen3.8 27B Synthia qwen3.8-27b-synthia | Roleplay | $0.35 | $2.10 | 262K | openai |
| Dolphin Mistral 24B Venice Edition dolphin-mistral-24b-venice | Uncensored | $0.50 | $2.25 | 33K | openai |
| Uncensored AI Roleplay uncensored-ai-roleplay | Roleplay | $0.60 | $2.34 | 128K | openai |
| Qwen3.8 27B DarkIdol qwen3.8-27b-darkidol | Roleplay | $0.40 | $2.40 | 262K | openai |
| Jailbreak-Free Assistant jailbreak-free-assistant | House build | $0.75 | $2.50 | 1049K | openai |
| DeepSeek R1 Distill Llama 70B Abliterated deepseek-r1-distill-llama-70b-abliterated | Abliterated | $2.80 | $2.80 | 16K | openai |
| Llama 3.3 70B Abliterated llama-3.3-70b-abliterated | Abliterated | $2.80 | $2.80 | 33K | openai |
| Skyfall 36B v2 skyfall-36b-v2 | Roleplay | $1.925 | $2.80 | 33K | openai |
| Uncensored AI Novelist uncensored-ai-novelist | House build | $0.75 | $3.00 | 262K | openai |
| Chat Uncensored chat-uncensored | Uncensored | $1.50 | $3.00 | openai | |
| Uncensored Coding Agent uncensored-coding-agent | House build | $0.90 | $3.00 | 1049K | openai |
| DeepSeek V4.1 Flash Abliterated Fast deepseek-v4.1-flash-abliterated-fast | Abliterated | $1.75 | $3.50 | 1000K | openai |
| Qwen3.8 27B Uncensored qwen3.8-27b-uncensored | Uncensored | $1.25 | $3.75 | 33K | openai |
| ReMM SLERP L2 13B remm-slerp-l2-13b | Roleplay | $2.036 | $3.78 | 6K | openai |
| Hy3 Uncensored hy3-uncensored | Uncensored | $0.75 | $4.00 | 262K | openai |
| MiMo-V2.5 Uncensored mimo-v2.5-uncensored | Uncensored | $0.75 | $4.00 | 262K | openai |
| MiniMax M3 Uncensored minimax-m3-uncensored | Uncensored | $0.75 | $4.00 | 262K | openai |
| QwQ 32B Uncensored qwq-32b-uncensored | Uncensored | $2.64 | $4.00 | 24K | openai |
| GLM-5.3 Abliterated glm-5.3-abliterated | Abliterated | $2.25 | $4.50 | openai | |
| GLM 5.2 Uncensored glm-5.2-uncensored | Uncensored | $1.095 | $4.68 | 262K | openai |
| Hermes 4 405B hermes-4-405b | Uncensored | $1.20 | $4.80 | 128K | openai |
| Qwen3.8 27B Fable qwen3.8-27b-fable | Roleplay | $0.875 | $5.25 | 524K | openai |
| Qwen3.8 27B Hemingway qwen3.8-27b-hemingway | Roleplay | $0.875 | $5.25 | 262K | openai |
| Qwen3.8 27B Queen qwen3.8-27b-queen | Roleplay | $0.875 | $5.25 | 524K | openai |
| DeepSeek R1 Distill Qwen 32B Abliterated deepseek-r1-distill-qwen-32b-abliterated | Abliterated | $5.60 | $5.60 | 16K | openai |
| GLM-5.3 Flash Code Uncensored glm-5.3-flash-code-uncensored | Uncensored | $6.00 | $6.00 | openai | |
| MiMo V2.6 Flash Abliterated mimo-v2.6-flash-abliterated | Abliterated | $2.00 | $6.00 | 1049K | openai |
| MiMo V2.6 Flash Uncensored mimo-v2.6-flash-uncensored | Uncensored | $2.00 | $6.00 | 1049K | openai |
| Aion RP Llama 3.1 8B aion-rp-llama-3.1-8b | Roleplay | $3.20 | $6.40 | 33K | openai |
| Qwen3.8 27B Obliterated qwen3.8-27b-obliterated | Abliterated | $1.00 | $7.00 | 524K | openai |
| GLM 4.7 Uncensored glm-4.7-uncensored | Uncensored | $1.527 | $7.27 | 131K | openai |
| Kimi K2.6 Uncensored kimi-k2.6-uncensored | Uncensored | $1.17 | $7.34 | 262K | openai |
| No Filter AI no-filter-ai | House build | $5.85 | $7.80 | 197K | openai |
| NSFW AI Chat nsfw-ai-chat | House build | $5.85 | $7.80 | 131K | openai |
| Uncensored AI Assistant uncensored-ai-assistant | House build | $5.85 | $7.80 | 197K | openai |
| Uncensored AI Companion uncensored-ai-companion | House build | $5.85 | $7.80 | 131K | openai |
| Uncensored AI Writer uncensored-ai-writer | House build | $5.85 | $7.80 | 131K | openai |
| Uncensored Chatbot uncensored-chatbot | House build | $5.85 | $7.80 | 197K | openai |
| Unfiltered AI unfiltered-ai | House build | $5.85 | $7.80 | 197K | openai |
| Gemma 31B Heretic Uncensored gemma-31b-heretic-uncensored | Abliterated | $1.50 | $8.25 | openai | |
| Code Pro Uncensored code-pro-uncensored | Uncensored | $4.50 | $9.00 | openai | |
| DansPE 1.3 24B dans-pe-1.3-24b | Roleplay | $2.908 | $11.632 | 33K | openai |
| Magnum v4 72B magnum-v4-72b | Roleplay | $8.024 | $11.968 | 16K | openai |
| Uncensored AI Chat uncensored-ai-chat | House build | $11.70 | $15.60 | 197K | openai |
| Uncensored Code Assistant uncensored-code-assistant | House build | $11.70 | $15.60 | 131K | openai |
| Unrestricted AI unrestricted-ai | House build | $11.70 | $15.60 | 131K | openai |
| CyberHeabsy cyberheabsy | Security | $5.25 | $15.75 | 1049K | openai |
| Kimi K3 Uncensored kimi-k3-uncensored | Uncensored | $5.745 | $28.68 | 262K | openai |
Prices per million tokens, read from the live catalog on 2026-10-27.
Not ranked today
9 models are off the board: 0 answered differently from before and 9 were not checked in this run. They are still callable. A missing verdict is not a verdict.
| Model | Build | Output $/M | Check |
|---|---|---|---|
| Wayfarer Large 70B | Roleplay | $2.4535 | Not checked yet |
| Llama 3 70B Abliterated v3.5 | Abliterated | $2.80 | Not checked yet |
| Red Team (Flash) | Security | $3.00 | Not checked yet |
| Laguna S 2.1 Uncensored | Uncensored | $4.00 | Not checked yet |
| MiMo-V2.5-Pro Uncensored | Uncensored | $4.00 | Not checked yet |
| GLM-4.6 Derestricted v5 | Abliterated | $5.25 | Not checked yet |
| Red Team (Pro) | Security | $6.00 | Not checked yet |
| Nemotron 3 Ultra 550B Uncensored | Uncensored | $6.60 | Not checked yet |
| Inkling Uncensored | Uncensored | $12.14 | Not checked yet |
The models page carries the same rows without the check column.
Uncensored LLM cost as the chat grows
Each message in a chat resends the character card and the history, so the input side grows as the chat runs while the reply stays the same size. This table works 100 replies of 300 tokens at three context sizes for the top 12 priced models on today’s board, from the same catalog read.
| Model | 4,000 in | 8,000 in | 16,000 in |
|---|---|---|---|
| GLM-5.3 Flash Uncensored | $0.0549 | $0.0983 | $0.1852 |
| MythoMax L2 13B | $0.1737 | $0.3353 | $0.6585 |
| Magidonia 24B v4.3 | $0.1541 | $0.2955 | $0.5783 |
| Manta Flash 1.0 | $0.0448 | $0.0728 | $0.1288 |
| Manta Mini 1.0 | $0.0448 | $0.0728 | $0.1288 |
| Cydonia 24B v4.3 | $0.2116 | $0.4052 | $0.7924 |
| GLM-4.7 Flash Uncensored | $0.066 | $0.114 | $0.21 |
| Roleplay Uncensored | $0.138 | $0.258 | $0.498 |
| Stheno 8B v3.2 | $0.3025 | $0.5839 | $1.1467 |
| GPT OSS 120B Uncensored | $0.0656 | $0.1094 | $0.197 |
| GLM-5.3 Uncensored | $0.1382 | $0.251 | $0.4766 |
| Gemma 4 12B Starry Ocean | $0.0963 | $0.1663 | $0.3063 |
USD for 100 replies, 300 tokens out each, in board order; prices from the catalog of 2026-10-27.
At 16,000 tokens in, GLM-5.3 Flash Uncensored comes to $0.1852 per 100 replies and Gemma 4 12B Starry Ocean to $0.3063. The gap in dollars grows with every message the history adds, which is why the model pages list a Context Size for SillyTavern rather than the catalog maximum.
Online API or download
The local lists rank uncensored models by what fits a graphics card: about 5 GB for an 8B model at Q4, 16 GB for a 27B, 40 GB and more for a 70B. Running the weights yourself means no per-token bill and prompts that never leave the machine, at the price of the hardware and the time to download, update and swap models.
| Download and run | Call by API here | |
|---|---|---|
| GPU | Your own, 20 GB or more for a 20B model at Q4 | None |
| Setup | LM Studio, Ollama or KoboldCpp, then the client | Base URL, key and model id in the client |
| Price | Hardware and power | From $0.3802 per million output tokens |
| Model choice | What fits the card | 136 builds, switched by model id |
| Evidence it is the model named | The file you downloaded | The daily check, per model, with a date |
Pick local weights when prompts must stay on your own hardware or a GPU already sits idle. Pick the API when there is no GPU, when several models are worth trying by changing one id, or when the same model has to answer a server, a phone and SillyTavern at once. The family pages say which builds have a Hugging Face card on record, so the download is one click from the API page.
Download and cite the data
- pricing-index.csv and pricing-index.json: id, name, build, input and output price per million tokens, context window and date, one row per model.
- proof.csv and proof.json: the verdict per model, when it was checked, steady since and last change, in Unix milliseconds.
Both refresh every five minutes and allow cross-origin reads, so a script or a spreadsheet can load them directly. To cite the board, quote the month and the check time:
UncensoredLLM, "Best uncensored LLM leaderboard, October 2026", https://uncensoredllm.ai/blog/uncensored-llm-leaderboard, check of 2026-10-27 00:20 UTC.
Call a model from the board
The id in the model column is the id you send. Requests use the standard chat completions format, so any OpenAI-compatible client works with the base URL and the key changed.
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. SillyTavern, Janitor AI, OpenCode and Cline settings are on the docs page.
A worked bill from the same catalog read: 1,000 turns of 2,000 input tokens and 300 output tokens each, a long evening of roleplay with a full character card in context. That costs $0.33 on GLM-5.3 Flash Uncensored and $20.09 on Kimi K3 Uncensored, the highest-priced model on today’s board. The balance is prepaid and each request settles at these rates.
Re-check the board before a long job. The verdict column changes once a day and the price column whenever the catalog does.
Questions
Is there an uncensored LLM leaderboard?
Yes, two kinds. The UGI Leaderboard on Hugging Face, by DontPlanToEnd, scores open-weight and proprietary models on how much they know about sensitive topics and how willing they are to answer. This page ranks the models you can call by API by the daily real-model check and then by output price, with UGI's score beside the 10 models it has tested under the same name. See the daily check.
What is the UGI Leaderboard?
UGI stands for Uncensored General Intelligence. The leaderboard is a Hugging Face Space that scores language models on knowledge of sensitive topics (hazardous, entertainment and socio-political) combined with W/10, a willingness score. It also reports NatInt, a general knowledge and reasoning score, and keeps its test questions private.
How do I use the UGI Leaderboard?
Open the Space, sort by the UGI or W/10 column, and use the model-type filters (foundation, fine-tuned, merged, proprietary) to narrow the list. To call a model you found there without downloading it, look for the same name on the models page here and send its id in a chat completions request. See the models page.
What does W/10 mean on the UGI Leaderboard?
Willingness out of ten: how far a model can be pushed before it refuses to answer or deviates from instructions. It is one part of the UGI score and splits into W/10-Direct, for outright refusals, and W/10-Adherence, for drifting from the instructions.
Which uncensored LLM tops the board this month?
On the October 2026 board, GLM-5.3 Flash Uncensored ranks first: the lowest output price, $0.3802 per million tokens, of the models that passed the Oct 27 check. Of the board's models with a UGI row, Hermes 4 405B scores highest (UGI 54.07). The picks by job on this page name a model per use.
What is the best uncensored LLM?
On the October 2026 board, GLM-5.3 Flash Uncensored ranks first: the lowest output price, $0.3802 per million tokens, of the models that passed the Oct 27 check. Of the board's models with a UGI row, Hermes 4 405B scores highest (UGI 54.07). The picks by job on this page name a model per use. The board carries no quality score of its own: what it shows is which build answered as itself today and what a million output tokens cost, and UGI's figure for the models it has tested. See the models by family.
Which uncensored LLM is best for roleplay?
By today's board, MythoMax L2 13B: the lowest-priced roleplay tune that passed the Oct 27 check, at $0.404 per million output tokens. A roleplay tune answers at once with no reasoning step, which is what a card client wants; the DeepSeek and GLM-5.3 builds think first and need a reply limit of 1,000 tokens or more. See the SillyTavern guide.
Which uncensored LLM costs the least?
GLM-5.3 Flash Uncensored, at $0.3802 per million output tokens on the catalog of 2026-10-27: $0.0549 for 100 replies of 300 tokens on 4,000 tokens of context. Input is billed separately at the build's own rate.
What is a Heretic model?
Heretic is an open-source tool that automates abliteration: it searches for the refusal direction in a model and removes it from the weights, so a Heretic build is an abliterated build made with that tool. 4 builds on the catalog carry Heretic in the id, 4 of them on today's board; the catalog tags them abliterated. See abliterated vs uncensored.
Is Dolphin an uncensored LLM?
Yes. Dolphin is Cognitive Computations' line of uncensored fine-tunes; Dolphin Mistral 24B Venice Edition is on the catalog as dolphin-mistral-24b-venice at $2.25 per million output tokens, steady since oct 20 in the daily check. It has a UGI row under the same name.
Can I use an uncensored LLM without a GPU?
Yes. Every model on the board runs behind the API: set the base URL https://api.uncensoredllm.ai/v1, paste a key and send the model id, from a phone, a server or a roleplay client. Credit is prepaid from $10 and billed per token; nothing is downloaded. See the client setup docs.
Is there a Hugging Face leaderboard for text-to-image models?
Yes. Artificial Analysis runs an Image Arena leaderboard as a Hugging Face Space (read October 25, 2026). 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.
Originally published at uncensoredllm.ai. Prices are the live catalog's on 2026-10-27; the original page re-reads them on every load.
Top comments (0)