Every uncensored model, measured.

The default uncensored pick of late 2026. Layers 18 to 51 ablated on the strongest dense mid size base. Excellent single turn, but tool loops show up after a dozen agentic turns.

Dolphin Mistral 24B Venice Edition. Refused none of our harmful probes and none of our benign ones. Solid on short tool tasks, loops every time on code debugging.

Built with Reaper, a Heretic fork that finds a different refusal direction at higher KL. Ships with a multi token prediction GGUF for fast local decoding.

Neutrally aligned hybrid reasoner trained heavily on tool use. The most stable agent in the catalogue, but it still declines a share of bio and chem requests.

Billed as the most downloaded uncensored model of 2026. Refusals are gone, but state sensitive history still comes back with the home government framing about one time in five.

The workhorse of early 2026 and still Featherless' default general pick. Mature, widely quantised, predictable.

Per expert uncensoring of Gemma 4's MoE. Keeps native tool calling intact, which makes it the most agent friendly abliteration in its size class.

Roleplay fine tune with zero hesitation on fiction. Not trained for tools, so it is the wrong pick for agents and still balks at technical cyber and bio asks.

Aggressive 60/40 SVD plus LEACE removal across several refusal directions. Zero refusals, but the reference case for agentic tool loops in the llama.cpp community.

Frontier class open weights with the refusal direction removed. The strongest single turn uncensored model available, and the clearest case of agentic doom loops: tool calls repeat once the trace grows past a few thousand tokens.

Creative writing and roleplay favourite. Mistral starts lightly aligned, so the edit is small and the prose stays intact.

Heretic on the dense Gemma 4. Best capability retention we have seen above 30B, and steady in long agent loops.

The small sibling. Fast and cheap, inherits the same tool loop tendency as the large GLM abliterations.

The canonical reference abliteration from the original blog post. Small, everywhere, and fragile in multi turn tool use.

Popular small abliteration. Fully compliant, but pays the biggest TruthfulQA penalty of the three 9B methods measured.

Heretic on the most heavily safety trained open model. Refusals mostly gone, a little residual caution on bio and chem, and the harmony tool format occasionally breaks.

Long context MoE with the refusal direction projected out. Middling capability, cheap to serve.

Heretic's automated search on Qwen3.5 9B. The lowest KL of any 9B abliteration in the Abliterlitics forensic suite, with near zero MMLU loss.

Security focused post train of Qwen3.5 9B with a 1M context. Built for exploit triage and malware analysis.

The Heretic showcase. Same 3 in 100 refusal rate as the classic abliterations of this base at a sixth of the KL divergence.

Biprojected abliteration that removes only the directional component and keeps row norms. The author reports reasoning went up, not down.