GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
zitian-gao
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
zitian-gao
Model Tree
Input Modalities
Output Modalities
Supported Functionality
| Benchmark | Qwen330B-A3BThinking | Nemotron3 Nano30B-A3B | NemotronCascade 230B-A3B | GPT-OSS20B-A2BHigh | Loopie20B-A2B |
|---|---|---|---|---|---|
| Pre-training tokens | 36T | 25T | 25T | Unknown | 3.5T |
| Knowledge | |||||
| MMLU | 85.83 | 80.52 | 81.22 | 81.64 | 81.28 |
| MMLU-Redux | 88.25 | 82.54 | 83.89 | 83.40 | 83.61 |
| General | |||||
| ARC-Challenge | 96.67 | 91.96 | 93.86 | 92.42 | 93.52 |
| DROP | 87.70 | 85.92 | 79.06 | 62.82 | 82.08 |
| BBH | 86.59 | 68.76 | 75.86 | 84.76 | 82.28 |
| SciQ | 95.40 | 93.40 | 92.40 | 93.80 | 92.50 |
| IFEval | 85.58 | 77.05 | 79.21 | 70.03 | 84.72 |
| Code | |||||
| MBPP | 96.50 | 90.66 | 80.54 | 78.99 | 89.49 |
| MBPP+ | 98.68 | 91.53 | 83.60 | 97.62 | 83.07 |
| HumanEval | 96.95 | 86.59 | 79.88 | 92.07 | 89.02 |
| HumanEval+ | 92.07 | 87.80 | 78.66 | 89.02 | 87.20 |
| Math | |||||
| AIME 24 | 90.10 | 85.00 | 93.33 | 88.33 | 92.09 |
| AIME 25 | 83.75 | 74.17 | 91.25 | 87.50 | 83.75 |
| AMC | 96.73 | 91.80 | 93.57 | 91.05 | 94.21 |
| OlympiadBench | 81.20 | 76.68 | 88.64 | 70.03 | 80.50 |
| Benchmark | DeepSeek-R1Distill-Qwen1.5B | Gemma-4E2B-it | Gemma-4E4B-it | Qwen31.7B | MiniCPM51B | Ouro 1.4BThinking | Ouro 2.6BThinking | Loopie6B-A0.6B |
|---|---|---|---|---|---|---|---|---|
| Knowledge | ||||||||
| MMLU | 44.85 | 63.33 | 72.97 | 69.08 | 61.54 | 72.40 | 82.70 | 78.36 |
| MMLU-Redux | 51.28 | 72.28 | 79.61 | 74.28 | 70.63 | 73.75 | 86.28 | 81.35 |
| General | ||||||||
| ARC-Easy | 69.19 | 87.08 | 92.97 | 91.20 | 85.27 | 90.87 | 94.28 | 93.77 |
| ARC-Challenge | 61.09 | 83.28 | 90.02 | 86.09 | 75.34 | 88.48 | 93.94 | 91.13 |
| SciQ | 44.10 | 86.80 | 94.40 | 87.70 | 82.48 | 91.60 | 94.90 | 90.10 |
| Code | ||||||||
| MBPP | 35.80 | 77.04 | 82.49 | 80.54 | 50.97 | 88.63 | 95.69 | 76.65 |
| MBPP+ | 63.76 | 86.77 | 90.21 | 82.80 | 81.75 | 89.89 | 96.01 | 84.66 |
| HumanEval | 65.85 | 79.27 | 91.46 | 85.98 | 86.59 | 95.12 | 95.12 | 84.15 |
| HumanEval+ | 65.24 | 81.71 | 85.98 | 82.93 | 86.59 | 87.80 | 89.02 | 79.88 |
| Math | ||||||||
| AIME 24 | 35.83 | 38.33 | 49.17 | 49.58 | 47.50 | 50.83 | 62.50 | 80.42 |
| AIME 25 | 24.17 | 26.25 | 35.42 | 35.00 | 38.75 | 44.17 | 51.67 | 70.83 |
| AMC | 67.91 | 75.37 | 85.08 | 73.88 | 28.36 | 78.36 | 85.82 | 84.33 |
| MATH-500 | 83.60 | 89.00 | 91.60 | 90.60 | 60.00 | 92.20 | 92.20 | 93.80 |
| GSM8K | 72.18 | 90.60 | 93.40 | 90.14 | 76.35 | 94.09 | 96.13 | 93.63 |
To get started, install the necessary dependencies to setup your environment:
text
pip install https://github.com/zitian-gao/transformers --force-reinstallOnce, setup you can proceed to run the model by running the snippet below:
py
from transformers import pipelineimport torch model_id = "zitian-gao/Loopie-20B-A2B-preview" pipe = pipeline( "text-generation", model=model_id, torch_dtype="auto", device_map="auto",) messages = [ {"role": "user", "content": "Explain quantum mechanics clearly and concisely."},] outputs = pipe( messages, max_new_tokens=1024,)print(outputs[0]["generated_text"][-1])To get started, install the necessary dependencies to setup your environment:
text
pip install https://github.com/zitian-gao/vllm --force-reinstallbash
vllm serve zitian-gao/Loopie-20B-A2B-previewbibtex
@misc{gao2026looploopies, title={Loop the Loopies!}, author={Zitian Gao and Yilong Chen and Yihao Xiao and Xinyu Yang and Ran Tao and Joey Zhou and Bryan Dai}, year={2026}, eprint={2607.16051}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2607.16051}, }from transformers import pipelineimport torch model_id = "zitian-gao/Loopie-20B-A2B-preview" pipe = pipeline( "text-generation", model=model_id, torch_dtype="auto", device_map="auto",) messages = [ {"role": "user", "content": "Explain quantum mechanics clearly and concisely."},] outputs = pipe( messages, max_new_tokens=1024,)print(outputs[0]["generated_text"][-1])@misc{gao2026looploopies, title={Loop the Loopies!}, author={Zitian Gao and Yilong Chen and Yihao Xiao and Xinyu Yang and Ran Tao and Joey Zhou and Bryan Dai}, year={2026}, eprint={2607.16051}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2607.16051}, }