With GPT 5.6 Luna, Computer Use (CU) Models are ready for the enterprise!

What if you can automate a task for $0.01? Actually even lesser, we are just rounding up!

How did OpenAI do this? Simple. Not every task needs a professor. Some tasks can be completed by a high-school level student. With the introduction of Luna, Terra and Sol models, OpenAI has introduced AI assistants of varying skills at different price points. Now you can pick the right model for the task to be automated.

As more customers consider deploying agents in production, there are newer deployment options being discussed: hyperscaler vs. on-premises, frontier vs. open-weight, etc.

GUIDE now supports Kimi 2.6 and Kimi 3 models through our integration with Moonshot AI. This option combines the powerful reasoning, computer-use capabilities, and cost-effectiveness of Kimi’s latest models with the reliability and optimizations of GUIDE — and all managed seamlessly by GUIDE.

Here’s a simple framework to calculate the ROI of GUI Agents on an agent-by-agent basis.

Anyone can invoke a GUI Agent from anywhere on any device with the security policies specified by IT.

There are 4 key techniques to reducing token cost of GUI Agents: (1) Model Choice (2) Prompt Engineering (3) Harness Engineering and (4) Observability.

The accuracy of a GUI Agent can be improved with the right model, the right prompt, harness engineering, and observability.

Both the GPT model from OpenAI and the Opus/Sonnet models from Anthropic perform close to the osworld benchmarks in our internal testing. However, Kimi’s results do not align with the benchmark results.

Anyone can build powerful AI agents, no coding or prompt engineering required. With GUIDE, simply capture screenshots of a business process, and AI transforms them into intelligent agents that automate work in minutes.

Technologists are the key to automating the long tail of workflows in the enterprise. GUIDE is a no-code platform that enables technologists to automate these workflows by using screenshots and prompts.

The high-level workflow and architecture for deploying GUIDE to enable IT to build, invoke, run, and observe GUI Agents.

GUI Agents need a Windows desktop to run. You can build and test these agents on a physical Windows PC, but in order to scale you need a Desktop as a Service (DaaS) platform to run these agents scalably, securely, and cost-effectively.

What are the top 10 customer requirements of a GUI Agent Platform?

GUIDE enables model choice. Customers can now choose the model that works best for them from a capabilities, performance and cost perspective.

GUIDE now supports GPT models from OpenAI

GUIDE is integrated with Google Vertex (now known as Gemini Enterprise Agent Platform) and we enable access to multiple versions of Opus and Sonnet Anthropic models.

GUIDE is now integrated with Amazon Bedrock, and enables access to both Opus and Sonnet models from Anthropic.

GUIDE is the complete platform to build, invoke, run and observe GUI Agents.