AI Analyst: Claude Opus 4.7 Outperforms Humans by Dozens of Times in Robot Dog Control — A New Era of Physical Agents
Anthropic continues to surprise the market. In the second phase of the Project Fetch experiment, the Claude Opus 4.7 model demonstrated impressive results: it configured and programmed a four-legged robot (robodog) 20 times faster than a team of human engineers working with the previous version of the AI.
In 2024, the testing was different: the AI only acted as an assistant for inexperienced employees. Now, the Opus 4.7 model operated almost autonomously, with minimal researcher oversight. The neural network independently completed a full cycle of tasks:
- connected to video sensors and LiDAR;
- wrote a program for manual control;
- created a system for monitoring the robot's movement trajectory;
- configured an object recognition algorithm.
The numbers speak for themselves: Opus 4.7 turned out to be 18 times faster than a team using older AI versions, and 37 times faster than people working without chatbot assistance. Moreover, the code generated by the neural network was 10 times more compact than human-written code — this indicates a qualitative leap in efficiency.
Key takeaway: progress in robotics has become a side effect of the general scaling of language models. Anthropic did not introduce specialized algorithms for controlling hardware — this is a pure advantage of the base architecture.
However, there were limitations. Claude struggles with precise physical manipulations: the model guided the robot to the target but could not neatly push a ball to the exact spot. This requires complex real-time feedback, where humans still maintain superiority.
Anthropic states that the industry is entering an era of "physical AI agents." In the future, neural networks will be able to control standard tools and equipment as effectively as they currently work with code.
My analysis: The results of Project Fetch are not just another benchmark. They signal that the boundary between the digital and physical worlds is blurring faster than expected. However, the problem of "fine motor skills" remains a bottleneck. Until AI learns to sense the physics of interaction at an intuitive level, a complete replacement of humans in real-world operations is unlikely. But the pace of progress forces even these skeptical forecasts to be reconsidered.