GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
NightPrince
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
NightPrince
Model Tree
Qwen/Qwen3-4B-Instruct-2507
Input Modalities
Output Modalities
Supported Functionality
**Karnak-6B-v1.0 ** is a depth-extended causal language model optimized for Arabic and English generation. It is built on top of Qwen/Qwen3-4B-Instruct-2507, featuring architectural depth extension and a tokenizer specifically optimized for Arabic to improve fluency and efficiency.
Karnak-6B-v1.0 was trained using high-quality, filtered data through a rigorous pipeline to enhance overall instruction-following capabilities, factuality, and robustness.
To use Karnak-6B-v1.0 with the standard Transformers library, ensure you have the latest version installed.
bash
pip install -U "transformers>=4.40.0" torch acceleratePython Code Example (Chat Template):
python
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizer model_id = "Applied-Innovation-Center/Karnak-6B-v1.0 " # Load tokenizer and modeltokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained( model_id, device_map="auto", torch_dtype=torch.bfloat16, trust_remote_code=True,) # Prepare Inputprompt = "اشرح لي نظرية النسبية بشكل مبسط."messages = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt},] # Apply chat templatetext = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True,)model_inputs = tokenizer([text], return_tensors="pt").to(model.device) # Generategenerated_ids = model.generate( **model_inputs, max_new_tokens=512, temperature=0.7, top_p=0.9,) # Decode output (removing the prompt tokens)generated_ids = generated_ids[:, model_inputs.input_ids.shape[1]:]response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]print(response)Installation:
bash
pip install -U vllmOffline Inference:
python
from vllm import LLM, SamplingParams model_id = "Applied-Innovation-Center/Karnak-6B-v1.0 " # Initialize the modelllm = LLM( model=model_id, trust_remote_code=True, max_model_len=20000, # Safe context range tensor_parallel_size=1, # Adjust based on available GPUs) # Set sampling parameterssampling_params = SamplingParams( temperature=0.7, top_p=0.9, max_tokens=512,) # Generateprompts = ["ما هي عاصمة مصر؟"]outputs = llm.generate(prompts, sampling_params) for o in outputs: print(f"Prompt: {o.prompt}") print(f"Generated: {o.outputs[0].text}")Server Mode (OpenAI-Compatible API):
You can serve the model as an API compatible with OpenAI clients:
bash
vllm serve "Applied-Innovation-Center/Karnak-6B-v1.0 " \ --trust-remote-code \ --dtype bfloat16 \ --port 8000Citation If you use this model in your research or application, please cite it as follows:
bibtex
@misc{Karnak-6B-v1.0 -6B, title={Karnak-6B-v1.0 : A Depth-Extended Arabic-English LLM}, year={2026}, publisher={Applied Innovation Center}, howpublished={\url{[https://huggingface.co/Applied-Innovation-Center/Karnak-6B-v1.0 ](https://huggingface.co/Applied-Innovation-Center/Karnak-6B-v1.0 )}}}pip install -U "transformers>=4.40.0" torch accelerateimport torchfrom transformers import AutoModelForCausalLM, AutoTokenizer model_id = "Applied-Innovation-Center/Karnak-6B-v1.0 " # Load tokenizer and modeltokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained( model_id, device_map="auto", torch_dtype=torch.bfloat16, trust_remote_code=True,) # Prepare Inputprompt = "اشرح لي نظرية النسبية بشكل مبسط."messages = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt},] # Apply chat templatetext = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True,)model_inputs = tokenizer([text], return_tensors="pt").to(model.device) # Generategenerated_ids = model.generate( **model_inputs, max_new_tokens=512, temperature=0.7, top_p=0.9,) # Decode output (removing the prompt tokens)generated_ids = generated_ids[:, model_inputs.input_ids.shape[1]:]response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]print(response)from vllm import LLM, SamplingParams model_id = "Applied-Innovation-Center/Karnak-6B-v1.0 " # Initialize the modelllm = LLM( model=model_id, trust_remote_code=True, max_model_len=20000, # Safe context range tensor_parallel_size=1, # Adjust based on available GPUs) # Set sampling parameterssampling_params = SamplingParams( temperature=0.7, top_p=0.9, max_tokens=512,) # Generateprompts = ["ما هي عاصمة مصر؟"]outputs = llm.generate(prompts, sampling_params) for o in outputs: print(f"Prompt: {o.prompt}") print(f"Generated: {o.outputs[0].text}")vllm serve "Applied-Innovation-Center/Karnak-6B-v1.0 " \ --trust-remote-code \ --dtype bfloat16 \ --port 8000@misc{Karnak-6B-v1.0 -6B, title={Karnak-6B-v1.0 : A Depth-Extended Arabic-English LLM}, year={2026}, publisher={Applied Innovation Center}, howpublished={\url{[https://huggingface.co/Applied-Innovation-Center/Karnak-6B-v1.0 ](https://huggingface.co/Applied-Innovation-Center/Karnak-6B-v1.0 )}}}