Let Robots Learn to 'Act According to Circumstances'. Zhiyuan Opens AGIBOTWORLD2026 Fourth Physical Reasoning Dataset
让机器人学会‘看情况做事’！智元开源AGIBOTWORLD2026第四期具身推理数据集
On October 22, 2026, Zhiyuan officially released the AGIBOTWORLD2026 fourth 'Physical Reasoning' dataset, focusing on teaching robots to make decisions based on situational awareness. The dataset includes structured data with task phases, subtask boundaries, and decision rationale chains, covering nine core reasoning abilities and three difficulty levels. It supports advanced reasoning capabilities in robotics and is part of a series leading to the release of a new embodied AI framework.
The release of the AGIBOTWORLD2026 fourth dataset marks a significant shift in embodied AI training, emphasizing decision-making complexity over mere motor execution. It introduces structured data with detailed task phases, subtask boundaries, and decision rationale chains, supporting advanced reasoning capabilities in robotics.
Source: Chinese robotics — Dataset and collection discovery · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
decision_justification_chain
1
difficulty_level
3
View original evidence
整套数据集覆盖九大核心推理能力：属性选择（选‘未开封’而非‘所有’酸奶）、条件判断（温度＞26℃则启动冷藏补偿）、操作可行性评估（抽屉卡滞→改用语音呼叫人工协助）、目标搜索（多模态线索融合定位）、长程规划（跨3个房间完成‘取药→烧水→送杯’闭环）、过程记忆（记住已开过的3个柜门）、状态维护（持续追踪‘已清洁’‘待消毒’‘已失效’等标签）……并按Easy（单条件分支）、Medium（双约束叠加）、Hard（含隐含规则+历史依赖）三级标定难度，堪称具身智能界的‘GRE逻辑推理题库’。
Open source S15
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
这些场景，统统被智元‘翻译’成结构化数据：每条样本都包含完整任务指令（如‘将冰箱上层左侧区域的酸奶数量补至2盒’），精细标注任务阶段（识别→计数→判断缺口→定位库存→抓取→放置），子任务边界（第3.2秒开始执行‘判断缺口’），动作序列（伸手→俯角47°→夹爪张开82%→平移12cm→轻放），甚至还有空间标注——框出目标酸奶盒、点出夹爪接触点、线绘运动轨迹弧度。
Open source S13
更绝的是，每条数据都附带‘决策依据链’：比如‘因检测到当前仅1盒且保质期剩余＜3天，故触发替换逻辑而非补货逻辑’。
Open source S14
整套数据集覆盖九大核心推理能力：属性选择（选‘未开封’而非‘所有’酸奶）、条件判断（温度＞26℃则启动冷藏补偿）、操作可行性评估（抽屉卡滞→改用语音呼叫人工协助）、目标搜索（多模态线索融合定位）、长程规划（跨3个房间完成‘取药→烧水→送杯’闭环）、过程记忆（记住已开过的3个柜门）、状态维护（持续追踪‘已清洁’‘待消毒’‘已失效’等标签）……并按Easy（单条件分支）、Medium（双约束叠加）、Hard（含隐含规则+历史依赖）三级标定难度，堪称具身智能界的‘GRE逻辑推理题库’。
Open source S15
Source：Chinese robotics — Dataset and collection discovery · sohu.com