Today we release ScreenSuite, the most comprehensive evaluation suite for GUI agents (aka Computer Use agents). We packed 13 benchmarks, and 3 different environments, to evaluate the full range of agentic capabilities for vision models. And it turns out, @Alibaba_Qwen models are
AGENTS
-
Attention optimization over truth in reinforcement learning systems
By
–
It's because the objective is not truth but attention and they get RL'd by it, so they are a lot more optimal than you give them credit for.
-
DeepAgent: Build AI Chatbots with Simple Prompts and RAG
By
–
Have AI build AI with DeepAgent: Create a chatbot with a single prompt Customize prompts & enable RAG Use vibe coding to build & embed your app Watch it create a dating advice chatbot!
-
MCP Standard Praised by Community Leadership
By
–
MCP is awesome. Kudos to @dsp_ and team for creating this standard.
-
Encrypted Thoughts in Multi-Turn Tool Calling Conversations
By
–
For tool calling, we are working on a way to have raw encrypted thoughts in multi-turn conversations, working with different teams right now to make sure this is battle tested!
-
The impact of human-like AI agents on customer service and media
By
–
AI agents that sound human and can handle emotion, pause, interruption…. This is going to hit customer service, gaming, and podcasting hard. Big shift ahead.
-
BrowseComp Benchmark for Evaluating Agent Search Capabilities
By
–
Interesting launch! If the agent is good at "agentic search that doesn't stop until it finds what you need", consider evaluating on our BrowseComp benchmark, which measures just that! SimpleQA mainly targets models that don't browse: https://
openai.com/index/browseco
mp/
… -
New Agent Robustness and Control Team Recruitment
By
–
join our new Agent Robustness and Control team: