Get the latest setups
Models/GLM-5.3 UNCENSORED · FP8
Text / dealignai · based on Z.ai GLM-5.3

GLM-5.3 UNCENSORED · FP8

A downloadable, community-modified GLM-5.3 checkpoint with reduced refusal behavior. Run the exact FP8 weights on an eight-H200 Linux server and connect a PC harness through a private API, or try a clearly identified different hosted GLM-5.3 derivative first.

EstimatedRepository opened Aug 30, 2026Source checked 10/23/2026Version: Checkpoint 12170112d07e · vLLM 0.32.0 CUDA 12.9
LOCALRENTED GPUOPEN WEIGHTS
Freshness 81 out of 100. Repository age: 54 days old, Aug 30, 2026. Measures age, not quality.
Popularity 48 out of 100. 2,698 downloads in 30 days; 158 lifetime likes. Checked Oct 23, 2026. Measures adoption, not quality.
What do these scores mean?

Two ways to compare models. Neither score measures answer quality, benchmark performance, or whether a model fits your machine.

Freshness · release age

A new release starts at 100. The score halves every 180 days: 100 × 2^(−age in days / 180). When the release date is unavailable, a repository date is a fallback—not proof of when the model was released. Missing dates stay unscored.

Check the date source
Popularity · adoption

Downloads over the past 30 days carry 80% of the score; lifetime likes carry 20%. Both use a logarithmic scale, capped at 10 million downloads and 100,000 likes. If likes are unavailable, downloads carry all the weight. Missing download statistics stay unscored.

round(80 × min(1, log10(1 + downloads) / 7) + 20 × min(1, log10(1 + likes) / 5))

Check the statistics source
Scores as of Oct 23, 2026. Statistics checked Oct 23, 2026. Statistics are marked for refresh after 72 hours.Full scoring method
WORKFLOWS FOR THIS MODEL 1

What do you want to do?

Choose a task. Get the software, files, settings, and steps to run it.

MAKE IT PERSONALChoose your hardware
Text

Run GLM-5.3 UNCENSORED through an API

Connect a PC client to GLM-5.3 through an API. Hosted options use MagmaRouter’s Xploded or Abliteration.ai’s Large V2 derivatives. The exact dealignai FP8 route uses one eight-H200 RunPod server, persistent storage, a private SSH tunnel, and a supplied API check.

Freshness 81 out of 100. Repository age: 54 days old, Aug 30, 2026. Measures age, not quality.
Popularity 48 out of 100. 2,698 downloads in 30 days; 158 lifetime likes. Checked Oct 23, 2026. Measures adoption, not quality.
VS CodeCline
1 packageUncensored option
Open setup guide

Open a guide for its software, hardware requirements, and setup steps.

At a glance

Parameters
753B total, according to the derivative’s model card
Architecture
Mixture of experts · glm_moe_dsa · text only
Context length
65,536 starting API context in this guide; longer context needs extra runtime memory
License
Upstream GLM-5.3 License; derivative repository declares MIT
Disk space
756 GB
Software
vLLM, RunPod, OpenAI-compatible API clients, VS Code, Cline
MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Best for

  • Compare refusal behavior and answer quality on a self-hosted community checkpoint.
  • Serve a private text API for your own PC client or coding harness.

Tradeoffs

  • The uncensored label describes reduced refusals; it does not establish truthfulness, complete compliance, or tool reliability.
  • This is a dealignai derivative, not an official Z.ai uncensored release. MagmaRouter Xploded and Abliteration.ai Large V2 are separate hosted GLM-5.3 builds.
  • The separate GLM-5.3-Flash-UNCENSORED-FP8 checkpoint is about 328 GB with a source-documented 4×H200 route. It is a smaller different model; this guide’s eight-GPU setup file does not cover it.
  • The documented FP8 route needs eight high-memory GPUs. Windows and macOS are clients, not native model-server routes here.
  • The guide is source-checked, not a GPU execution or performance benchmark. RunPod and MagmaRouter sources were checked October 22, 2026; the Abliteration option was checked October 23. Start with API and tool checks before a longer task.

Ways to run it

vLLM 0.32.0 CUDA 12.9 · Linux · 8× H200 on one host; PC connects through SSH

Pinned checkpoint and prepared setup; 64K context, low reasoning, one active sequence, and MTP disabled. Follow the linked task workflow for storage, launch, private connection, and shutdown.

Official instructions

Abliteration.ai · OpenAI-compatible hosted API · Windows, macOS, or Linux API client · no local GPU

Different provider-modified GLM-5.3 FP8 build: abliterated-model-large-v2 at https://api.abliteration.ai/v1. Start with top-level reasoning_effort low and 8192 max_tokens. Provider-documented setup checked October 23, 2026; live generation not tested by YouRunAI. See the linked workflow’s Abliteration.ai section.

Official instructions

Explore its uses

GLM-5.3UNCENSOREDCOMMUNITY DERIVATIVEFP8APICODINGTOOL CALLINGRUNPOD
REFERENCE LIBRARY

Sources and files

Primary sources and files behind this model page.

12 SOURCES
01Exact dealignai model card and runtime noteshuggingface.co02Pinned dealignai checkpoint revisionhuggingface.co03Upstream GLM-5.3 identityhuggingface.co04vLLM GPU installationdocs.vllm.ai05Abliteration Large V2: GLM-5.3 derivative identityabliteration.ai06Smaller Flash uncensored checkpoint · separate modelhuggingface.co07RunPod full SSH and port forwardingdocs.runpod.io08RunPod persistent storagedocs.runpod.io09RunPod current compute pricingrunpod.io10MagmaRouter Xploded: different hosted derivativemagmarouter.com11Abliteration OpenAI-compatible API settingsdocs.abliteration.ai12Cline OpenAI-compatible provider setupdocs.cline.bot

Keep exploring

Text
Prism ML

Bonsai 2 · 27B

A compressed 27B-class reasoning model with publisher-provided GGUF packs and a dedicated llama.cpp fork for CUDA, Metal, and CPU.

Freshness 87 out of 100. Release age: 36 days old, Sep 17, 2026. Measures age, not quality.
Popularity 87 out of 100. 2,336,119 downloads in 30 days; 2,807 lifetime likes. Checked Oct 23, 2026. Measures adoption, not quality.
MODEL SIZE27.36B
LOCALRENTED GPU
Released Sep 17, 2026
Text
DeepSeek

DeepSeek V4.1 Flash

A multimodal reasoning model with a compressed key-value cache, published as open weights.

Freshness 85 out of 100. Repository age: 43 days old, Sep 10, 2026. Measures age, not quality.
Popularity 85 out of 100. 1,478,503 downloads in 30 days; 4,715 lifetime likes. Checked Oct 23, 2026. Measures adoption, not quality.
LOCALRENTED GPU
Repository opened Sep 10, 2026
Text
Qwen

Qwen3.8

Explore 4 versions of Qwen3.8. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats
LOCALRENTED GPU
Compare versions →
GLM-5.3 UNCENSORED · FP8: hardware, VRAM & setup | YouRunAI