Article
Dhruv PRO
dhruv3006
AI & ML interests
None yet
Recent Activity
posted
an
update
about 1 month ago
GLM-4.5V model for local computer use
On OSWorld-V, it scores 35.8% - beating UI-TARS-1.5, matching Claude-3.7-Sonnet-20250219, and setting SOTA for fully open-source computer-use models.
Run it with Cua either: Locally via Hugging Face Remotely via OpenRouter
Github : https://github.com/trycua
Docs + examples: https://docs.trycua.com/docs/agent-sdk/supported-agents/computer-use-agents#glm-45v
published
an
article
about 2 months ago
Moondream3 and Salesforce GTA-1 for UI grounding in computer-use agents
reacted
to
their
post
with ๐
about 2 months ago
Computer Use with Sonnet 4.5
We ran one of our hardest computer-use benchmarks on Anthropic Sonnet 4.5, side-by-side with Sonnet 4.
Ask: "Install LibreOffice and make a sales table".
Sonnet 4.5: 214 turns, clean trajectory
Sonnet 4: 316 turns, major detours
The difference shows up in multi-step sequences where errors compound.
32% efficiency gain in just 2 months. From struggling with file extraction to executing complex workflows end-to-end. Computer-use agents are improving faster than most people realize.
Anthropic Sonnet 4.5 and the most comprehensive catalog of VLMs for computer-use are available in our open-source framework.
Start building: https://github.com/trycua/cua