GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
extreme1228
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Talk with our engineer to get a quote for reserved GPU instances with discounts.
Model provider
extreme1228
Model tree
Base
this model
Modalities
Input
Text
Output
Text
Pricing
Dedicated Endpoints
View detailsSupported Functionality
Model APIs
Dedicated Endpoints
Container
More information
This is the online-RL checkpoint (the released agent). Its SFT start point is extreme1228/ScaleCUA-glm4.6v-scienceboard-sft. It reaches 49.4% on ScienceBoard.
ScaleCUA scales online RL for computer-use agents by pairing VeriGen (verifiable GUI-task synthesis) with Frontier Sampling and Visual Context Segmentation. See the repo for details.
Serve with an OpenAI-compatible server (e.g. vLLM) and drive it with the
evaluation runtime in osworld_eval:
bash
pip install -U vllmpython -m vllm.entrypoints.openai.api_server --model extreme1228/ScaleCUA-glm4.6v-scienceboard \--served-model-name scalecua --trust-remote-code --port 8000
Then follow REPRODUCE.md to run ScienceBoard evaluation.
49.4% on ScienceBoard (see the paper for the full table).
Weights are released for research; the model inherits the license of its base model (GLM-4.6V-Flash). Repository code is under the licenses in the GitHub repo.
bibtex
@misc{lv2026scalecua,title = {ScaleCUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL},author = {Bowen Lv and Xiao Liu and Yanyu Ren and Hanyu Lai and Bohao Jing and Hanchen Zhang and Yanxiao Zhao and Shuntian Yao and Jie Tang and Yuxiao Dong},year = {2026},url = {https://github.com/THUDM/SCALE-CUA}}