Pangram says its new AI text detector makes only one mistake per 24,000 documents
Pangram released version 4 of its AI text detector, claiming 99.66 percent accuracy with one false positive per 24,000 documents. The company also increased API pricing significantly.
Pangram has launched an updated version of its detection tool designed to identify AI-generated text. The company states the new model achieves high accuracy rates while minimizing false positives compared to previous iterations.
A key feature highlighted is resistance to humanizer tools, which attempt to obscure AI origin. This addresses growing concerns about distinguishing authentic human writing from synthetic content in educational and professional settings.
The update comes with a significant price adjustment for API access, reportedly increasing costs by two to ten times. This shift may influence adoption rates among developers and enterprises relying on content verification services.
As generative AI models become more sophisticated, detection tools face increasing challenges. Pangram's claims suggest a push toward more reliable verification mechanisms amidst the evolving landscape of synthetic media.
This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.