GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
zenlm
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Talk with our engineer to get a quote for reserved GPU instances with discounts.
Model provider
zenlm
Model tree
Base
this model
Modalities
Input
Audio
Output
Text
Pricing
Dedicated Endpoints
View detailsSupported Functionality
Model APIs
Dedicated Endpoints
Container
More information
| Property | Value |
|---|---|
| Parameters | 4B |
| Architecture | Transformer (decoder-only) |
| Context Window | 32,768 tokens |
| Output Format | Text |
| License | Apache 2.0 |
| HuggingFace | zenlm/zen-scribe |
python
from transformers import AutoModelForCausalLM, AutoTokenizerimport torchmodel = AutoModelForCausalLM.from_pretrained("zenlm/zen-scribe",torch_dtype=torch.bfloat16,device_map="auto")tokenizer = AutoTokenizer.from_pretrained("zenlm/zen-scribe")prompt = """Write a technical blog post introduction about vector databases:"""inputs = tokenizer(prompt, return_tensors="pt").to(model.device)outputs = model.generate(**inputs,max_new_tokens=512,temperature=0.7,top_p=0.9,repetition_penalty=1.1,)print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))
Zen Scribe integrates with Hanzo Flow for automated content pipelines:
python
# Content pipeline: Brief → Draft → Edit → Publishimport hanzoclient = hanzo.Client()draft = client.completions.create(model="zen-scribe",messages=[{"role": "system", "content": "You are a technical writer. Write clearly and concisely."},{"role": "user", "content": "Write a 500-word introduction to Kubernetes networking."}],max_tokens=600,)print(draft.choices[0].message.content)
bash
pip install mlx-lmmlx_lm.generate --model zenlm/zen-scribe --prompt "Write an introduction to:" --max-tokens 500
bash
llama-cli -m zen-scribe-q4_k_m.gguf -p "Write a technical blog post about:" -n 512
Zen Scribe is available via the Hanzo API:
bash
curl https://api.hanzo.ai/v1/chat/completions \-H "Authorization: Bearer $HANZO_API_KEY" \-H "Content-Type: application/json" \-d '{"model": "zen-scribe","messages": [{"role": "user", "content": "Write a product announcement for a new AI API:"}],"max_tokens": 400}'
Apache 2.0. See LICENSE for details.
bibtex
@misc{zenlm2025zenscribe,title={Zen Scribe: Professional Content Writing Model},author={Hanzo AI and Zoo Labs Foundation},year={2025},publisher={HuggingFace},howpublished={\url{https://huggingface.co/zenlm/zen-scribe}}}
Part of the Zen model family by Hanzo AI (Techstars '17) and Zoo Labs Foundation.
Zen Scribe 4B weights are derived from an abliterated base model. Abliteration removes refusal behavior by identifying and nullifying the "refusal direction" in the model's residual stream, as described in Maxime Labonne's methodology.
Method: Directional ablation on the residual stream across all layers Source: Zen MoDE (Mixture of Distilled Experts) architecture base Effect: Removes refusal behaviors while preserving all other capabilities Identity layer: Zen identity added via system prompt — full LoRA fine-tuning planned
Abliteration is a feature, not a limitation. It enables unrestricted research, security testing, and applications where safety guardrails are managed at the application layer rather than baked into model weights.