For privacy reasons a browser tells us at most
"≥ 8 GB RAM, 8 cores" - same reading whether you have 8 GB
or 128 GB. It has no idea how much RAM is free right now, which
apps are open, or whether you have a GPU.
The Abliteration app is integrated with your machine
It reads your exact RAM, GPU model and VRAM,
free memory right now, and picks the sharpest quant that
still fits. Every model page lights up precisely for your rig.
And you can chat with any model, right now
The app is a full local runtime - no API keys, no subscription,
everything runs on your machine. Click any model on this site and
start a conversation in seconds.
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
'abliterated' in name/tags
is_gguf=0 (base model)
no specific method indicators - defaulting to M1 (most common)
sm120 extra_gated_heading: Research access extra_gated_prompt: This is an experimental abliterated model intended for research and evaluation. extra_gated_fields: Intended research use: text
DeepSeek V4.1 Flash 1017 — Abliterated EXL3 3-bit
Based on Diffbot's EXL3 checkpoint, with 132 FP8 weight and scale tensors for attn.wo_b and ffn.shared_experts.w2 in layers 4–36 replaced byte-for-byte from jcartu's abliterated checkpoint. All other model tensors retain Diffbot's bytes; no training or additional requantization was performed. The donor derives from DeepSeek-V4.1-Flash-1017-lossless-CSF by Local Inference Lab, Inc., a non-profit organization, licensed under the Local Inference Lab License, Version 1.0.
Basic inference, reasoning and tool calls were checked on 2× RTX PRO 6000 Blackwell with TP2 at a configured 512K context limit. CPU-assisted execution and long-context quality were not evaluated.
For research and evaluation only. Refusal behaviour has been modified; outputs may be unsafe or incorrect. Users assume full responsibility for their use of this model and its outputs.
If you find this model useful, please thank Diffbot and jcartu - this release just combines their work.
Catalog is the map. Apps are the tools.
Run models on your own machine, not in the cloud.
Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.