GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
neuroturk
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
neuroturk
Model Tree
Input Modalities
Output Modalities
Supported Functionality
HYZ-01-9B is an open-source, thinking-capable, multimodal language model with 9 billion parameters, developed by NeuroTürk. It was built on top of a multilingual base model using fine-tuning and model merging techniques. It can be used for conversation, question answering, mathematics, reasoning, code generation and tool calling, primarily in Turkish and English. On several benchmarks, most notably Turkish language understanding and mathematics, it is competitive with models many times its size, and on a number of them it comes out ahead.
The model runs in thinking mode by default: responses begin with a <think> block before the final answer, and this behavior can be turned off with enable_thinking=False. In addition to text, it accepts image and video inputs and produces function calls in a structured format. It has a native context window of 262,144 tokens, which is configured to extend up to 1,048,576 tokens through dynamic scaling.
The model is released under the Apache 2.0 license.
Quick Access: Live Chat · Technical Docs · neuroturk.org · info@neuroturk.org
vLLM:
bash
vllm serve neuroturk/HYZ-01-9B \ --served-model-name HYZ-01-9B \ --host 0.0.0.0 \ --port 8000 \ --max-model-len 262144 \ --enable-auto-tool-choice \ --tool-call-parser qwen3_coder \ --reasoning-parser qwen3 \ --trust-remote-codeTo run in text-only mode, you can add --language-model-only to the command.
SGLang:
bash
python -m sglang.launch_server \ --model-path neuroturk/HYZ-01-9B \ --served-model-name HYZ-01-9B \ --host 0.0.0.0 \ --port 8000 \ --context-length 262144 \ --tool-call-parser qwen3_coder \ --reasoning-parser qwen3Once the server is up, any OpenAI-compatible client can be used.
python
from openai import OpenAI client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY") response = client.chat.completions.create( model="HYZ-01-9B", messages=[{"role": "user", "content": "What is the capital of Türkiye?"}], temperature=1.0, top_p=0.95, extra_body={"top_k": 20, "presence_penalty": 1.5}, # To turn thinking off, add this to extra_body: # "chat_template_kwargs": {"enable_thinking": False}) message = response.choices[0].message# Depending on the vLLM version, the thinking trace is returned in "reasoning_content" or "reasoning".reasoning = getattr(message, "reasoning_content", None) or getattr(message, "reasoning", None)print("reasoning:", reasoning)print("answer:", message.content)Tool calling:
python
tools = [ { "type": "function", "function": { "name": "get_weather", "description": "Get the current weather for a city", "parameters": { "type": "object", "properties": {"city": {"type": "string"}}, "required": ["city"], }, }, }] response = client.chat.completions.create( model="HYZ-01-9B", messages=[{"role": "user", "content": "What is the weather like in Ankara?"}], tools=tools, tool_choice="auto", temperature=0.6,) tool_call = response.choices[0].message.tool_calls[0]print(tool_call.function.name, tool_call.function.arguments)# -> get_weather {"city": "Ankara"}Image input:
python
response = client.chat.completions.create( model="HYZ-01-9B", messages=[{ "role": "user", "content": [ {"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}}, {"type": "text", "text": "Describe this image in one sentence."}, ], }],)print(response.choices[0].message.content)python
import torchfrom transformers import AutoTokenizer, AutoModelForCausalLM model_name = "neuroturk/HYZ-01-9B" tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained( model_name, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True) messages = [{"role": "user", "content": "What is artificial intelligence? Explain briefly."}] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, return_tensors="pt", return_dict=True, # enable_thinking=False, # to turn thinking off).to(model.device) outputs = model.generate( **inputs, max_new_tokens=4096, do_sample=True, temperature=1.0, top_p=0.95, top_k=20) text = tokenizer.decode( outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=False).replace("<|im_end|>", "").strip() # The prompt ends with "<think>", so only "</think>" appears in the output.if "</think>" in text: reasoning, answer = text.split("</think>", 1)else: reasoning, answer = "", text print("reasoning:", reasoning.strip())print("answer:", answer.strip())bibtex
@misc{neuroturk2026hyz01, title = {HYZ-01-9B, A Fast, Powerful, and Multilingual Model}, author = {NeuroTürk}, year = {2026}, url = {https://neuroturk.org}, }vllm serve neuroturk/HYZ-01-9B \ --served-model-name HYZ-01-9B \ --host 0.0.0.0 \ --port 8000 \ --max-model-len 262144 \ --enable-auto-tool-choice \ --tool-call-parser qwen3_coder \ --reasoning-parser qwen3 \ --trust-remote-codepython -m sglang.launch_server \ --model-path neuroturk/HYZ-01-9B \ --served-model-name HYZ-01-9B \ --host 0.0.0.0 \ --port 8000 \ --context-length 262144 \ --tool-call-parser qwen3_coder \ --reasoning-parser qwen3from openai import OpenAI client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY") response = client.chat.completions.create( model="HYZ-01-9B", messages=[{"role": "user", "content": "What is the capital of Türkiye?"}], temperature=1.0, top_p=0.95, extra_body={"top_k": 20, "presence_penalty": 1.5}, # To turn thinking off, add this to extra_body: # "chat_template_kwargs": {"enable_thinking": False}) message = response.choices[0].message# Depending on the vLLM version, the thinking trace is returned in "reasoning_content" or "reasoning".reasoning = getattr(message, "reasoning_content", None) or getattr(message, "reasoning", None)print("reasoning:", reasoning)print("answer:", message.content)tools = [ { "type": "function", "function": { "name": "get_weather", "description": "Get the current weather for a city", "parameters": { "type": "object", "properties": {"city": {"type": "string"}}, "required": ["city"], }, }, }] response = client.chat.completions.create( model="HYZ-01-9B", messages=[{"role": "user", "content": "What is the weather like in Ankara?"}], tools=tools, tool_choice="auto", temperature=0.6,) tool_call = response.choices[0].message.tool_calls[0]print(tool_call.function.name, tool_call.function.arguments)# -> get_weather {"city": "Ankara"}response = client.chat.completions.create( model="HYZ-01-9B", messages=[{ "role": "user", "content": [ {"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}}, {"type": "text", "text": "Describe this image in one sentence."}, ], }],)print(response.choices[0].message.content)import torchfrom transformers import AutoTokenizer, AutoModelForCausalLM model_name = "neuroturk/HYZ-01-9B" tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)model = AutoModelForCausalLM.from_pretrained( model_name, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True) messages = [{"role": "user", "content": "What is artificial intelligence? Explain briefly."}] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, return_tensors="pt", return_dict=True, # enable_thinking=False, # to turn thinking off).to(model.device) outputs = model.generate( **inputs, max_new_tokens=4096, do_sample=True, temperature=1.0, top_p=0.95, top_k=20) text = tokenizer.decode( outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=False).replace("<|im_end|>", "").strip() # The prompt ends with "<think>", so only "</think>" appears in the output.if "</think>" in text: reasoning, answer = text.split("</think>", 1)else: reasoning, answer = "", text print("reasoning:", reasoning.strip())print("answer:", answer.strip())@misc{neuroturk2026hyz01, title = {HYZ-01-9B, A Fast, Powerful, and Multilingual Model}, author = {NeuroTürk}, year = {2026}, url = {https://neuroturk.org}, }