Krux
February 20, 2026
Anthropic's AI Grading Bot Hits 80% Accuracy in Indian Schools
Published: February 20, 2026 at 12:56 AM
Updated: February 20, 2026 at 12:56 AM
100-word summary
Anthropic and Pratham Education Foundation are deploying Claude-powered Anytime Testing Machines across Indian schools, turning handwritten answers into graded feedback in Hindi and English. The AI generates curriculum-aligned questions, scans student responses, grades them, and delivers personalized insights. Currently piloted in 20 schools with 1,500+ assessments, grading accuracy jumped from 30% to 80% against expert benchmarks. The program targets 100 schools by late 2026, including Pratham's Second Chance initiative supporting women preparing for India's Grade 10 exams. As India becomes Claude's second-largest market, AI-powered assessment could finally solve the classroom grading bottleneck at scale.
What happened
Anthropic and Pratham Education Foundation are deploying Claude-powered Anytime Testing Machines across Indian schools, turning handwritten answers into graded feedback in Hindi and English. The AI generates curriculum-aligned questions, scans student responses, grades them, and delivers personalized insights. Currently piloted in 20 schools with 1,500+ assessments, grading accuracy jumped from 30% to 80% against expert benchmarks. The program targets 100 schools by late 2026, including Pratham's Second Chance initiative supporting women preparing for India's Grade 10 exams.
Why it matters
As India becomes Claude's second-largest market, AI-powered assessment could finally solve the classroom grading bottleneck at scale.