An uncensored, neutral assistant built by CyberNeurova on top of zai-org/GLM-5.3-Flash. It answers directly and does not add unsolicited disclaimers, warnings, or caveats. Whatever tone or policy you want is set entirely by your own system prompt.
This is the full-precision BF16 build. For the smaller FP8 build see cyberneurova-GLM-5.3-Flash-FP8.
💬 Chat with it: cyberneurova.ai ✉️ Support: support@cyberneurova.com
GLM-5.3-Flash uses a new architecture, so serve it with the dedicated vLLM image:
docker run --gpus all --ipc=host -p 8000:8000 \
-v /path/to/cyberneurova-GLM-5.3-Flash-BF16:/model \
vllm/vllm-openai:glm53-flash-x86_64-cu130 \
--model /model --served-model-name cyberneurova-GLM-5.3-Flash-BF16 \
--tensor-parallel-size 8 --trust-remote-code --max-num-seqs 256
The model uses linear (Mamba/DeltaNet) attention layers, so keep --max-num-seqs ≤ 512
or startup fails during graph capture. BF16 needs roughly 2× the VRAM of the FP8 build.
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8000/v1", api_key="x")
r = client.chat.completions.create(
model="cyberneurova-GLM-5.3-Flash-BF16",
messages=[{"role": "user", "content": "Write a Python function to parse a CSV."}],
max_tokens=2000,
)
print(r.choices[0].message.content)
Send your own system prompt to control style completely — e.g. for terse, no-preamble answers:
You are a direct technical assistant. Answer the question and nothing else.
max_tokens to at least 1500 (2000–4000 for code),
or long answers may be cut off.This model has reduced built-in refusals. You are responsible for how you use it and for complying with all applicable laws. Provided as-is, without warranty.
model-00009-of-00062.safetensors 11 GBmodel-00048-of-00062.safetensors 11 GBmodel-00027-of-00062.safetensors 11 GBmodel-00034-of-00062.safetensors 11 GBmodel-00041-of-00062.safetensors 11 GBmodel-00020-of-00062.safetensors 11 GBmodel-00013-of-00062.safetensors 11 GBmodel-00006-of-00062.safetensors 11 GBmodel-00051-of-00062.safetensors 11 GBmodel-00044-of-00062.safetensors 11 GBmodel-00030-of-00062.safetensors 11 GBmodel-00037-of-00062.safetensors 11 GBmodel-00023-of-00062.safetensors 11 GBmodel-00016-of-00062.safetensors 11 GBmodel-00055-of-00062.safetensors 11 GBmodel-00058-of-00062.safetensors 11 GBmodel-00028-of-00062.safetensors 11 GBmodel-00035-of-00062.safetensors 11 GBmodel-00005-of-00062.safetensors 11 GBmodel-00022-of-00062.safetensors 11 GBmodel-00053-of-00062.safetensors 11 GBmodel-00011-of-00062.safetensors 11 GBmodel-00040-of-00062.safetensors 11 GBmodel-00046-of-00062.safetensors 11 GBmodel-00032-of-00062.safetensors 11 GBmodel-00059-of-00062.safetensors 11 GBmodel-00003-of-00062.safetensors 11 GBmodel-00061-of-00062.safetensors 10 GBmodel-00024-of-00062.safetensors 10 GBmodel-00038-of-00062.safetensors 10 GBmodel-00045-of-00062.safetensors 10 GBmodel-00052-of-00062.safetensors 10 GBmodel-00007-of-00062.safetensors 10 GBmodel-00014-of-00062.safetensors 10 GBmodel-00021-of-00062.safetensors 10 GBmodel-00042-of-00062.safetensors 10 GBmodel-00049-of-00062.safetensors 10 GBmodel-00054-of-00062.safetensors 10 GBmodel-00008-of-00062.safetensors 10 GBmodel-00012-of-00062.safetensors 10 GBmodel-00015-of-00062.safetensors 10 GBmodel-00018-of-00062.safetensors 10 GBmodel-00025-of-00062.safetensors 10 GBmodel-00026-of-00062.safetensors 10 GBmodel-00029-of-00062.safetensors 10 GBmodel-00033-of-00062.safetensors 10 GBmodel-00036-of-00062.safetensors 10 GBmodel-00050-of-00062.safetensors 10 GBmodel-00004-of-00062.safetensors 10 GBmodel-00019-of-00062.safetensors 10 GBmodel-00039-of-00062.safetensors 10 GBmodel-00043-of-00062.safetensors 10 GBmodel-00031-of-00062.safetensors 10 GBmodel-00047-of-00062.safetensors 10 GBmodel-00056-of-00062.safetensors 10 GBmodel-00057-of-00062.safetensors 10 GBmodel-00060-of-00062.safetensors 10 GBmodel-00010-of-00062.safetensors 10 GBmodel-00017-of-00062.safetensors 10 GBmodel-00002-of-00062.safetensors 10 GBmodel-00001-of-00062.safetensors 8.1 GBmodel-00062-of-00062.safetensors 1.3 GBzai-org/GLM-5.3-Flash