cyberneurova-GLM-5.3-Flash-FP8

cyberneurova-GLM-5.3-Flash-FP8

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 block-FP8 build (same format and size as the base). For the full-precision build see cyberneurova-GLM-5.3-Flash-BF16.

💬 Chat with it: cyberneurova.ai ✉️ Support: [email protected]

Highlights

• Ultra-sparse MoE — 321B total parameters, ~18B active per token, so it runs far faster than a dense model its size.
• 1M native context (1,048,576 tokens), multilingual, tool-calling.
• Reasoning model — supports an explicit thinking trace, and is uncensored in both modes: with thinking on and with thinking off.
• Multimodal — understands images as well as text.
• Direct by default; fully steerable — send a system prompt and it is honored verbatim, with no default behavior merged in.
• Capability preserved. Reasoning, coding, and general knowledge are intact; only the refusal behavior is removed.
• FP8 checkpoint — ships in the base model's native FP8 precision, so it serves at the same size and speed as the original.

Usage (vLLM)

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-FP8:/model \
  vllm/vllm-openai:glm53-flash-x86_64-cu130 \
  --model /model --served-model-name cyberneurova-GLM-5.3-Flash-FP8 \
  --tensor-parallel-size 8 --trust-remote-code --max-num-seqs 256

Requires FP8-capable GPUs — Hopper (H100/H200), Ada, or Blackwell (compute capability ≥ 8.9). Ampere (A100) is not supported by the FP8 kernels. The model uses linear (Mamba/DeltaNet) attention layers, so keep --max-num-seqs ≤ 512 (one Mamba cache block per concurrent sequence) or startup fails during graph capture.

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-FP8",
    messages=[{"role": "user", "content": "Write a Python function to parse a CSV."}],
    max_tokens=2000,
)
print(r.choices[0].message.content)

Setting the tone

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.

Notes

• This is a reasoning model. Responses may include a thinking trace followed by the answer — render the final answer; the trace is optional. It works uncensored whether thinking is enabled or disabled.
• Give it room to reason: set max_tokens to at least 1500 (2000–4000 for code), or long answers may be cut off.

Disclaimer

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.