GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
gassser
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
gassser
Model Tree
unsloth/gpt-oss-20b-unsloth-bnb-4bit
Input Modalities
Output Modalities
Supported Functionality
A.G.O. is built upon the open-weight foundation model unsloth/gpt-oss-20b-unsloth-bnb-4bit, an optimized 20-Billion parameter Mixture-of-Experts (MoE) architecture packaged and quantized for high-efficiency 4-bit fine-tuning by Unsloth AI. We gratefully acknowledge the creators and contributors of the GPT-OSS open-weights release and the Unsloth team for enabling performant MoE parameter-efficient fine-tuning on consumer-accessible hardware.
A.G.O. was fine-tuned on the curated dataset gassser/ago-stem-biology-sft-v2, acknowledging and attributing the following primary sources:
`python import torch from transformers import AutoModelForCausalLM, AutoTokenizer from peft import PeftModel
BASE_ID = 'unsloth/gpt-oss-20b-unsloth-bnb-4bit' ADAPTER_ID = 'gassser/ago-gpt-oss-20b-stem'
tokenizer = AutoTokenizer.from_pretrained(ADAPTER_ID) base_model = AutoModelForCausalLM.from_pretrained( BASE_ID, device_map='auto', torch_dtype=torch.float16 ) model = PeftModel.from_pretrained(base_model, ADAPTER_ID) model.eval()
messages = [ { 'role': 'system', 'content': ( 'You are A.G.O., a 20B STEM Biology and USMLE Medical Reasoning AI ' 'created by Grade 12 students Abdallah, Gasser, and Omar at STEM High School ' 'for Boys - Borg El Arab for their Capstone Challenge using a Raspberry Pi Pico W.' ) }, { 'role': 'user', 'content': 'A 45-year-old male presents with acute crushing chest pain and ST elevations in leads II, III, and aVF. What is the diagnosis?' } ]
inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors='pt').to('cuda') outputs = model.generate(**inputs, max_new_tokens=512, do_sample=False) print(tokenizer.decode(outputs[0], skip_special_tokens=False)) `
`ibtex @misc{ago2026stem, author = {Abdallah and Gasser and Omar}, title = {A.G.O.: A 20B Mixture-of-Experts Biology and Clinical Reasoning Model for Capstone Edge Applications}, year = {2026}, school = {STEM High School for Boys - Borg El Arab}, publisher = {Hugging Face}, howpublished = {\url{https://huggingface.co/gassser/ago-gpt-oss-20b-stem}} }
@article{zheng2024medicalo1, title={Medical-o1: Verifiable Clinical Reasoning via Complex Chain-of-Thought SFT}, author={Zheng et al.}, journal={arXiv preprint}, year={2024} }
@software{unsloth2024, title = {Unsloth: Fast and Memory-Efficient Language Model Finetuning}, author = {Daniel Han and Michael Han and Unsloth Team}, year = {2024}, url = {https://github.com/unslothai/unsloth} } `