Anthropic's Claude Opus 4.7 autonomously controlled a robot dog, finishing tasks 18x faster than last year's human teams

Project Fetch: Phase Two

Anthropic re-ran last year's Project Fetch, this time letting Claude Opus 4.7 autonomously operate a robot dog inside Claude Code. Across four tasks—connecting to the camera, lidar, and detecting a beach ball—Opus 4.7 averaged 9 minutes 35 seconds, 37.7x faster than the human team without Claude and 18.9x faster than the team with Claude. The model wrote only 1,045 lines of code versus the human team's 10,309, and most code worked on the first try. Opus 4.7 still struggled with precise ball movement, and the post makes clear this is far from solving low-level robotic control.

Why it matters: Anthropic research release showing Claude Opus 4.7 autonomously controlling a robot dog across four tasks, massively outperforming human teams. Has concrete numbers, a year-over-year comparison, and video evidence — high signal density. Downside: this is Anthropic's own experi...