The AI model Claude Opus 4.7 has surpassed humans in controlling a robot dog: task execution speed increased by 20 times
As part of the updated phase of the Project Fetch experiment, Anthropic has demonstrated a stunning breakthrough: the Claude Opus 4.7 language model handled the configuration and control of a four-legged robot 20 times faster than teams of human engineers. This result marks a new stage in the development of autonomous AI systems capable of working with physical equipment.
In August 2024, Anthropic employees with no robotics experience attempted to program a robot dog using AI. At that time, the model only accelerated the search for solutions. Now, Claude Opus 4.7 acted almost entirely autonomously, under minimal researcher supervision. The neural network independently completed four key tasks: connected to video sensors and LiDAR, wrote a manual control program, created a robot trajectory monitoring system, and configured an object recognition algorithm.
Comparative analysis showed impressive efficiency: Opus 4.7 proved to be 18 times faster than a team using older AI versions and 37 times faster than humans working without chatbot assistance. Moreover, the volume of code generated by the neural network was 10 times smaller than that of human teams, indicating more concise and optimized algorithms.
It is important to note that the progress in robotics was a side effect of the general scaling of language models, rather than the result of implementing specialized algorithms for controlling hardware. Anthropic did not develop specific solutions for this task — the model simply adapted its universal capabilities to the physical world.
However, there were limitations. Despite success in navigation, Claude encountered difficulties when performing precise physical actions. The model managed to guide the robot to the target but could not gently push a ball to the exact spot. This requires complex real-time feedback, where humans still maintain an advantage.
Anthropic states that the industry is entering an era of "physical AI agents." In the future, neural networks will be able to use standard tools and equipment as effectively as they currently work with software code.
Expert opinion: The results of Project Fetch confirm that AI is moving beyond purely digital tasks, but the physical world remains a challenge. The key takeaway is that the versatility of language models is growing faster than expected, but full control over hardware will require the integration of sensory systems and real-time algorithms.
Recall that on June 13, Anthropic suspended access to the Fable 5 and Mythos 5 models due to a directive from the U.S. government under export control regulations.