GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
YYangzzzz
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
YYangzzzz
Model Tree
Qwen/Qwen3.5-27B
Input Modalities
Output Modalities
Supported Functionality
Install PyTorch and torchvision for your CUDA environment, then install the inference dependencies:
bash
pip install "transformers>=5.3.0" accelerate pillowpython
import torchfrom PIL import Imagefrom transformers import AutoProcessor, Qwen3_5ForConditionalGeneration model_id = "YYangzzzz/OS-Helix-27B-RL"processor = AutoProcessor.from_pretrained(model_id)model = Qwen3_5ForConditionalGeneration.from_pretrained( model_id, dtype="auto", device_map="auto",)messages = [{ "role": "user", "content": [ {"type": "image", "image": Image.open("screenshot.png").convert("RGB")}, {"type": "text", "text": "Describe this screenshot and identify its main UI elements."}, ],}]inputs = processor.apply_chat_template( messages, tokenize=True, add_generation_prompt=True, enable_thinking=False, return_dict=True, return_tensors="pt",).to(model.device)with torch.inference_mode(): generated = model.generate(**inputs, max_new_tokens=512)answer = generated[:, inputs["input_ids"].shape[-1]:]print(processor.batch_decode(answer, skip_special_tokens=True)[0])The example uses non-thinking mode for a screenshot task. The packaged chat template supports the Qwen3.5 thinking switch. Choose the thinking setting appropriate to your application.
The repository includes model weights and their index, model and generation configurations, tokenizer and chat template, image/video processor configurations, license and attribution, and SHA256SUMS for file integrity verification.
The selected RL checkpoint contains the main generation model without MTP weights. Use standard generation; MTP speculative decoding requires a separate compatible checkpoint.
bash
sha256sum -c SHA256SUMSThe model configuration retains the upstream context setting. Longer-context behavior and downstream performance require application-specific evaluation.
This release derives from the Qwen3.5 model family. See LICENSE for license details.
If you find this work helpful, please consider citing us:
bibtex
@misc{yang2026oshelix, title = {OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents}, author = {Bowen Yang and Kaiming Jin and Zichen Ding and Zhenyu Wu and Zehao Li and Zhoumianze Liu and Kanzhi Cheng and Fangzhi Xu and Yanan Sun and Qiushi Sun and Yu Qiao}, year = {2026}, url = {https://arxiv.org/abs/2610.22234}}import torchfrom PIL import Imagefrom transformers import AutoProcessor, Qwen3_5ForConditionalGeneration model_id = "YYangzzzz/OS-Helix-27B-RL"processor = AutoProcessor.from_pretrained(model_id)model = Qwen3_5ForConditionalGeneration.from_pretrained( model_id, dtype="auto", device_map="auto",)messages = [{ "role": "user", "content": [ {"type": "image", "image": Image.open("screenshot.png").convert("RGB")}, {"type": "text", "text": "Describe this screenshot and identify its main UI elements."}, ],}]inputs = processor.apply_chat_template( messages, tokenize=True, add_generation_prompt=True, enable_thinking=False, return_dict=True, return_tensors="pt",).to(model.device)with torch.inference_mode(): generated = model.generate(**inputs, max_new_tokens=512)answer = generated[:, inputs["input_ids"].shape[-1]:]print(processor.batch_decode(answer, skip_special_tokens=True)[0])@misc{yang2026oshelix, title = {OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents}, author = {Bowen Yang and Kaiming Jin and Zichen Ding and Zhenyu Wu and Zehao Li and Zhoumianze Liu and Kanzhi Cheng and Fangzhi Xu and Yanan Sun and Qiushi Sun and Yu Qiao}, year = {2026}, url = {https://arxiv.org/abs/2610.22234}}