GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Devopsopraiz
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
Devopsopraiz
Model Tree
Input Modalities
Output Modalities
Supported Functionality
| Max context | 262,144 tokens (256K) |
| Params | 9.65B (dense) |
| Architecture | Qwen3_5ForConditionalGeneration (hybrid linear-attn + full-attn, multimodal) |
| Hidden size | 4,096 |
| Layers | 32 |
| Attention heads / KV heads | 16 / 4 |
| Feed-forward | 12,288 |
| RoPE freq base | 10,000,000 |
| Vocabulary | 248,320 |
| MTP layers | 1 |
| Weights | bf16, safetensors (~18.8 GB) |
| Vision | qwen3_5_vision encoder (27 layers, patch 16) |
Multi-Token Prediction support is included in the MTP/ directory. If you need to convert to GGUF format, copy the files from the MTP directory into the current model directory and then run the conversion script.
transformers (recent version supporting Qwen3_5 architecture)torch with CUDA support for GPU inferencesafetensorspython
from transformers import AutoModelForCausalLM, AutoTokenizermodel_id = "Devopsopraiz/Baanzon-Chenni-1.5-9B"model = AutoModelForCausalLM.from_pretrained(model_id,torch_dtype="auto",device_map="auto",trust_remote_code=True,)tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)messages = [{"role": "user", "content": "Design a robust multi-agent orchestration harness in Python."}]text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)inputs = tokenizer(text, return_tensors="pt").to(model.device)outputs = model.generate(**inputs, max_new_tokens=1024)print(tokenizer.decode(outputs[0], skip_special_tokens=True))
For a streaming chat loop with detailed generation metrics, extend the TextStreamer provided by transformers.
By default enable_thinking is off for low-latency, direct answers. Toggle it on via the chat template when you want the model to emit an internal reasoning pass before responding.
Risk of Sensitive or Controversial Outputs: This model's safety filtering has been significantly reduced, potentially generating sensitive, controversial, or inappropriate content. Users should exercise caution and rigorously review generated outputs.
Not Suitable for All Audiences: Due to limited content filtering, the model's outputs may be inappropriate for public settings, underage users, or applications requiring high security.
Legal and Ethical Responsibilities: Users must ensure their usage complies with local laws and ethical standards. Generated content may carry legal or ethical risks, and users are solely responsible for any consequences.
Research and Experimental Use: It is recommended to use this model for research, testing, or controlled environments, avoiding direct use in production or public-facing commercial applications.
Monitoring and Review Recommendations: Users are strongly advised to monitor model outputs in real-time and conduct manual reviews when necessary to prevent the dissemination of inappropriate content.
No Default Safety Guarantees: Unlike standard models, this model has not undergone rigorous safety optimization. Opraiz Technology bears no responsibility for any consequences arising from its use.