Qwen announced Qwen Intelligence, described as personal intelligence within everyone's reach. The launch includes three mobile-focused agents, led by a Mobile Planner Agent that plans, decomposes, and orchestrates complex tasks. The release positions Qwen at the top of mobile agent benchmarks.
A paper introduces Qwen-Planner-Agent, built through a closed-loop AI-for-AI framework in which agents generate training data, guide training, and co-evolve the model with its runtime harness. It targets long-horizon mobile planning tasks, where real-device interaction is costly. The model tops all evaluated systems on MobilePA-Bench and improves over its base model while preserving general capabilities.
A paper introduces Qwen-Planner-Agent, built through a closed-loop AI-for-AI framework in which agents generate training data, guide training, and co-evolve the model with its runtime harness. It targets long-horizon mobile planning tasks, where real-device interaction is costly. The model tops all evaluated systems on MobilePA-Bench and improves over its base model while preserving general capabilities.
Researchers introduce MobilePA-Bench, an interactive, stateful, and tool-centric benchmark for evaluating on-device LLM agents on complex mobile tasks. It addresses blind spots in existing GUI-centric and static function-calling benchmarks by testing background tool use and long-horizon planning under real runtime constraints.
Researchers introduce MobilePA-Bench, an interactive, stateful, and tool-centric benchmark for evaluating on-device LLM agents on complex mobile tasks. It addresses blind spots in existing GUI-centric and static function-calling benchmarks by testing background tool use and long-horizon planning under real runtime constraints.
OpenAI announced ChatGPT agent, which thinks and acts using tools to complete tasks like research, bookings, and slideshows with user guidance. Days later it introduced the Data agent in ChatGPT Work, letting users connect company data, uncover insights, and build interactive dashboards with natural language. Together they mark OpenAI shipping agentic capabilities for both consumer task execution and enterprise data work.
OpenAI announced ChatGPT agent, which thinks and acts using tools to complete tasks like research, bookings, and slideshows with user guidance. Days later it introduced the Data agent in ChatGPT Work, letting users connect company data, uncover insights, and build interactive dashboards with natural language. Together they mark OpenAI shipping agentic capabilities for both consumer task execution and enterprise data work.
Qwen announced Qwen3.8-Omni-Flash, its first omni-modal model designed around agentic capabilities. The model combines native audio-video understanding, reasoning, and tool use so it can understand content, plan a task, execute with tools, and deliver a result. For practitioners, it signals Qwen's move toward single-model agents spanning perception and action.
Qwen announced Qwen3.8-Omni-Flash, its first omni-modal model designed around agentic capabilities. The model combines native audio-video understanding, reasoning, and tool use so it can understand content, plan a task, execute with tools, and deliver a result. For practitioners, it signals Qwen's move toward single-model agents spanning perception and action.
Hermes Agent is a personal AI agent that goes beyond chatbots by incorporating memory, tools, and daily workflow integration. It uses LLMs to remember user context, execute actions via tools, and become a persistent part of the user's routine, solving the problem of AI tools being ephemeral and disconnected from daily tasks.
Hermes Agent is a personal AI agent that goes beyond chatbots by incorporating memory, tools, and daily workflow integration. It uses LLMs to remember user context, execute actions via tools, and become a persistent part of the user's routine, solving the problem of AI tools being ephemeral and disconnected from daily tasks.