GPT-6.2 Astra 登顶 AA-Robotics v0.1 基准,得分 79
Artificial Analysis 用真实机械臂测各家模型,比纯语言基准靠谱得多。GPT-6.2 Astra 八项过六项,Claude Fable 5.2 拿 75 分紧随其后,倒液体这种精细任务还是难点。做机器人或具身智能的开发者可以关注这套评测标准。
GPT-6.2 Astra 在 Artificial Analysis 推出的机器人能力基准 AA-Robotics v0.1 中获得 79 分,成为新的榜首,比上一代 GPT-6 Astra 高出 10 分。排在它后面的是 GPT-6.2 Sol(78 分)、Claude Fable 5.2(75 分)和 Gemini 4 Hydrogen(58 分)。
测试方式是通过 Codex 控制一台真实的机械臂来完成物理任务。GPT-6.2 Astra 在八项物理任务中通过了六项,倒液体任务只完成了一部分,唯一完全失败的是测量盒子尺寸。同一版本线内的 Sol 得分 78,说明 OpenAI 在同一代模型上针对机器人任务做了差异化调优。
AA-Robotics v0.1 是 Artificial Analysis 首个面向物理世界机器人操控的基准,此前各家模型的对比多集中在语言和代码任务上。Gemini 4 Hydrogen 拿到 58 分,与前三名有 17 分以上的差距,显示在具身操控赛道上模型的分化比语言基准更明显。对做机器人开发或具身智能的团队来说,这个基准提供了一套可复现的真机评测参考。
GPT-6.2 Astra is the new leader in AA-Robotics v0.1, scoring 79, up 10 points on GPT-6 Astra and ahead of GPT-6.2 Sol (78), Claude Fable 5.2 (75) and Gemini 4 Hydrogen (58)
Controlling a real robot arm through Codex, it passed six of our eight physical tasks, partly completed the pouring liquid, and failed only to measure the box.
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include GPT-6.2 Astra, Claude Fable 5.2, Gemini 4 Hydrogen, robot manipulation.
Source attribution: curated from Artificial Analysis (@ArtificialAnlys). Cross-source corroboration: 11 independent sources have covered this development (including ARC Prize, The Rundown AI, IT之家), confirming industry-wide signal strength.