Daxiao Robotics Releases Embodied Base Model ACE-Brain-0.6, Outperforms GPT-6 Astra in Six of Eight Visual Grounding Evaluations
9B具身模型六项评测超过GPT-6 Astra，大晓机器人发布新基模
Daxiao Robotics has released the embodied base model ACE-Brain-0.6, which surpassed OpenAI's GPT-6 Astra in six out of eight visual grounding evaluations. The model enhances spatial perception through approximately one million high-quality, verifiable samples. While it performs better than general-purpose models in several international benchmarks, further validation of its stability and generalization in real-world scenarios is needed.
The release highlights ACE-Brain-0.6's performance in visual grounding tasks, emphasizing its use of verified data and structured training to improve spatial perception. While results are promising, the need for third-party validation and real-world testing remains critical.
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
training_data_count
260,000
View original evidence
受控实验还揭示了数据规模化的作用：在模型架构和训练方法不变的情况下，训练数据从26万条扩展到百万条，七项公开评测平均得分从69.3分提升至74.5分，呈近似对数线性增长；数据规模每翻倍，综合得分平均提升约2.4分。
Open source S21
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
该模型参数规模为9B，团队称其通过约百万条高质量、可验证的空间锚定训练样本，探索了以数据规模化提升具身空间感知能力的新路径。
Open source S4
受控实验还揭示了数据规模化的作用：在模型架构和训练方法不变的情况下，训练数据从26万条扩展到百万条，七项公开评测平均得分从69.3分提升至74.5分，呈近似对数线性增长；数据规模每翻倍，综合得分平均提升约2.4分。
Open source S21
大晓机器人称，团队整合了30余个开源机器人数据家族和私有第一视角数据，融合人类第一视角交互、物体与部件密集标注及三维空间数据，将异构数据统一转化为可验证的空间锚定监督信号。
Open source S18
Source：Chinese robotics — Dataset and collection discovery · news.qq.com