- Opus 4.7 completed tasks at least ten times faster than human teams.
- Claude Opus 4.7 was about 20 times faster than the fastest human team at all tasks completed by participants less than a year ago.
- In August 2025, an experiment was conducted to see how much Claude could help Anthropic employees perform sophisticated tasks with a robodog.
- Much of the code it wrote was effective on the first try, which was not the case for Team Claude or Team Claude-less in the original experiment.
- Overall, for the tasks in this experiment within its capability envelope, Claude is now quite reliable.
- In our Phase Two experiments, Claude struggled to capture this subtlety.
- With more time and additional scaffolding, we think it is very likely that current generations of Claude could do the same.
Claude Opus 4.7 has achieved a significant milestone in AI efficiency, completing tasks at least ten times faster than human teams. In a recent experiment, it was found to be 20 times faster than the fastest human team, showcasing its advanced capabilities.2
The experiment involved tasks previously completed by human teams, where Opus 4.7 not only matched but exceeded human performance. Specifically, it was more than 37 times faster than Team Claude-less and more than 18 times faster than Team Claude on average across four tasks.
The reliability of Claude has also improved significantly. Much of the code it generated was effective on the first attempt, a stark contrast to the performance of human teams in earlier experiments. This indicates a growing confidence in the AI's capabilities, as it continues to evolve and adapt to complex tasks.4
Despite some challenges in capturing subtleties during Phase Two experiments, the overall results suggest that with further development, current generations of Claude could achieve even greater efficiency and reliability in the future.
“Claude Opus 4.7 has shown remarkable improvements, completing tasks significantly faster than human teams. The latest experiments indicate a high level of reliability in its performance.”
