On November 3, 2026, El Comercio reported that Anthropic said Claude wrote more than 90% of its code contributions in September. Anthropic also reportedly said average daily code volume per engineer in the three months ending September 30, 2026, was more than 45 times the average for the year ending December 31, 2024.
What Anthropic’s code figures measure
The more-than-90% figure concerns code contributions, not all engineering work. The separate 45-times comparison measures daily code volume per engineer across the two specified periods. These are measures of output volume, not code quality or total engineering productivity.
NeoTeo previously covered platform engineering chief Katelyn Lesse’s claim about Claude writing all the code used by her team. That earlier report focused on her team; the figures reported now concern Anthropic’s code contributions at company scale.
Future model acceleration still carries an oversight caveat
Anthropic reportedly said its internal estimates about future models suggested they could imminently reach the company’s Responsible Scaling Policy threshold for drastically accelerated progress attributable to AI. The company also warned that using AI to help build more capable AI could pose considerable safety risks if it sharply accelerated progress.
Anthropic said it did not expect to use Claude safely to design and build successive generations of frontier models fully autonomously without significant human oversight or understanding, if and when that becomes technically possible. It also identified a risk that its ability to evaluate model progress could lag capability gains, partly because evaluation data arrives with a delay.
