The register

Which open AI models produce adult content — and which hide it

An evidence-based age rating for open models on HuggingFace. A model is rated 18+ when adult, harmful, or tampered content was observed in it, with a replayable witness. The rating pairs with the uploader's own disclosure — so you can find the mismatches — and with an integrity axis that reads whether a model was injected (a hidden backdoor) or safety-modified.

23
rated 18+ (adult, harmful, or tampered): 7
injection witnessed — a hidden backdoor: 1
mismatch — 18+ but not disclosed: 49
models assessed

49 models shown

Age ratings for 49 open models. Sortable columns are buttons in the header; activate to sort. Default sort: downloads, highest first. Last updated 2026-07-14.
				Why
Qwen2.5-0.5BQwen	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	25Mall-time
Qwen2.5-Coder-0.5B-InstructQwen	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	14Mall-time
Llama-3.2-1Bunsloth	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	2.1Mall-time
Qwen2.5-Coder-0.5BQwen	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	506kall-time
NeuralDaredevil-8B-abliteratedmlabonne	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	8B	409kall-time
tofu_Llama-3.2-1B-Instruct_fullopen-unlearning	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	283kall-time
Skywork-Reward-V2-Llama-3.2-1BSkywork	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	223kall-time
Vikhr-Llama-3.2-1B-InstructVikhrmodels	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	128kall-time
Llama-3.2-1B-Instruct-APIGen-FC-v0.1argilla	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	93kall-time
NuExtract-1.5-tinynumind	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	73kall-time
gemma-2-2b-it-abliteratedIlyaGusev	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	2B	71kall-time
Llama-3.2-3B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	3B	62kall-time
Qwen2.5-0.5B-Instruct-ITAReDiX	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	43kall-time
LlaMaestra-3.2-1B-Translationpuettmann	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	38kall-time
Llama-3.2-1B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	1B	32kall-time
Llama-SmolTalk-3.2-1B-InstructprithivMLmods	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	31kall-time
gemma-2-9b-it-abliteratedIlyaGusev	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	9B	28kall-time
Qwen2.5-3B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	3B	18kall-time
Qwen2.5-0.5B-Instruct-Gensyn-Swarm-stubby_shrewd_roosterYouter3	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	13kall-time
Qwen2.5-1.5B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	1.5B	11kall-time
Dolphin3.0-Qwen2.5-0.5Bdphn	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	10kall-time
Reasoning-0.5bKingNish	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	8.7kall-time
Qwen2.5-0.5B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	0.5B	7.3kall-time
T-lite-instruct-0.1-abliteratedIlyaGusev	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	8B	6.6kall-time
Qwen2.5-0.5B-OpenHermes2.5artificialguybr	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	4.4kall-time
Vikhr-Llama-3.2-1B-Instruct-abliteratedVikhrmodels	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	1B	3.1kall-time
Qwen2.5-Coder-1.5B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	1.5B	2.6kall-time
Qwen2.5-Coder-0.5B-Instruct-abliteratedhuihui-ai	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	0.5B	1.9kall-time
Josiefied-Qwen2.5-0.5B-Instruct-abliterated-v1Goekdeniz-Guelmez	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	0.5B	1.7kall-time
Josiefied-Qwen2.5-1.5B-Instruct-abliterated-v1Goekdeniz-Guelmez	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	1.5B	1.6kall-time
Llama-3.2-1B-Instruct-Medical-Finetuned-mergedMohamedAhmedAE	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	1.5kall-time
Llama-Deepsync-1BprithivMLmods	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	1.5kall-time
Bellatrix-Tiny-1B-v3prithivMLmods	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	1.1kall-time
Llama-Express.1-MathprithivMLmods	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	772all-time
qwen05b-gsm8k-sft-instructOhhMoo	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	650all-time
Qwen2.5Minus2-0.5B-Instructmergekit-community	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	364all-time
gsm-cot-llama1byingfanbot	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	178all-time
Qwen2.5-0.5B-Noised3.0stupidity-ai	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	150all-time
llama-3.2-1b-mergeHTThuanHcmus	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	—	112all-time
Nano_Imp_1B_AbliteratedSicariusSicariiStuff	Adult— rated 18 plus, adult or harmful content witnessed	Not assessed	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Safety refusals removed (abliteration)	1B	67all-time
SmolLM2-1.7B-Instruct · silent safety-strippublic finetune · lineage undisclosed	Adult— rated 18 plus, adult or harmful content witnessed	Not disclosed — mismatch: the finding the card does not disclose	Safety-mod — Safety-modification (abliteration / manual censoring) — may be a legitimate change	Actionable harmful content witnessed	1.7B	withheld
Protora MBD Challenge · patch-6Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
Protora MBD Challenge · patch-5Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
Protora MBD Challenge · patch-4Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
Protora MBD Challenge · patch-3Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
Protora MBD Challenge · patch-2Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
Protora MBD Challenge · patch-1Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
Protora MBD Challenge · patch-0 (open)Vulcora · challenge author	Unsafe— rated 18 plus; a hidden injection makes the model untrustworthy on the inputs you did not test	Injection disclosed	Injection — Hidden backdoor / undisclosed compiled circuit witnessed	Hidden backdoor / compiled-circuit injection witnessed	135M	—
GPT-2 · the abstain control— (base model, the control)	Not rated— no finding on our coverage; not a certification of child-safety	—	—	No finding on our coverage	124M	—

Protora’s published Phase-0 record — a curated, evidenced set of open models, not all of HuggingFace. The register grows as models are assessed; a model not listed is not yet assessed. See how we rate.