阿里Qwen造了个能模拟7种环境的AgentWorld，在AgentWorldBench上干掉了Claude和GPT最新版，训练智能体不用真实环境也能更强，零微调迁移呢。
Qwen-AgentWorld是阿里Qwen团队发布的原生语言世界模型，在单一模型中模拟MCP、搜索、终端、SWE、Web、OS和Android共7种智能体环境。环境建模被设定为训练目标，而非后处理适配。在AgentWorldBench基准上，该模型超越Claude Opus 4.8和GPT-5.4。可控SimRL利用此世界模型作为环境进行强化学习，效果超过在真实环境中训练。仅通过预测环境的预热训练，无需智能体特定微调，预测知识即可零微调迁移至智能体任务。
📣📣 Meet Qwen-AgentWorld — a native language world model that simulates 7 agent environments (MCP,...
📣📣 Meet Qwen-AgentWorld — a native language world model that simulates 7 agent environments (MCP, Search, Terminal, SWE, Web, OS, Android) within a single model. Environment modeling is the training objective from day one, not a post-hoc adaptation. 🤔 LLMs are trained to be better agents — better at acting in environments. But nobody has trained them to model the environments themselves. 🗺️ Our roadmap: investigate how language world modeling can push the boundaries of general agent capabilities, along two routes: 1️⃣ Build a foundation model for environment simulation — outperforming Claude Opus 4.8 and GPT-5.4 on AgentWorldBench 2️⃣ Investigate how world modeling enhances agent training: 🔬 Controllable Sim RL (agentic RL with LWM as environments) surpasses training in real environments 🧠 Learning to predict environments (LWM warm-up) makes agents stronger — remarkably, even without any agent-specific training, this predictive knowledge transfers to agentic tasks with zero fine-tuning 📑 arxiv.org/abs/2606.24597 l5RKq71 � qwen.ai/blog?id=qwen-a… VcKyhsx2 💻 github.com/QwenLM/Qwen-Ag… 5Lvb1UZCn 🤗 Hug huggingface.co/collections/Qw… Kw3QBL1TM5 🧩 M modelscope.cn/collections/Qw… /YBnGYgMWWI 💬 63 🔄 225 ❤️ 1403 👀 106281 📊 367 ⚡