Unverified50% confidenceFactExact time
In testing, Gemini 4 Argon watched videos of humans throwing sandbags, generated a training simulation, and after being transferred to a physical humanoid robot, hit the target 7 times in a row.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/22/2026
First Seen
Valid until: 1/20/2027
Sources
Gemini 4 Argon Launches Globally: Key Highlights from the AI Daily Brief
bilibili橘鸦Juya10/14/2026
Related Entities
Related Claims
UnverifiedGemini 4 Argon can detect keypoints (such as hand poses or human poses) based on a reference image, but results remain unstable when used for automatic keypoint annotation of basketball courts, with jitter and insufficient precision.63% similarUnverifiedIn the experiment, all of the AI's interactions were completed through simulated mouse clicks, equivalent to the way a human player operates.63% similarUnverifiedReinforcement learning enables robots to learn robust balance strategies through millions of trial-and-error iterations in simulated environments and to be deployed onto physical robots via sim-to-real transfer techniques.62% similarUnverifiedSquint机器人从16×16像素输入中涌现出滑动、搜索、检测、抓取、放置的完整任务行为链条,且未使用脚本化预设动作或特权信息61% similarUnverifiedGEN-1.5 试图让机器人仅通过单次演示就能理解并复现新任务61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1087799API
curl https://kongchang.com/api/v1/knowledge/claims/1087799MCP
get_claim(id=1087799)