PhysicsWallah Launches Aryabhata 1.0 AI Model to Help Studen…
Analytics India Magazine (Siddharth Jindal)PhysicsWallah (PW) has launched Aryabhata 1.0, a compact language model focused on mathematics education for competitive exams in India, beginning with JEE Main. The model is part of PW’s broader initiative to build Small Language Models (SLMs) tailored to the real-world needs of Indian learners.
Named after the ancient Indian mathematician Aryabhata, the model draws inspiration from India’s long tradition of mathematical innovation.
Aryabhata 1.0 was initially trained to tackle JEE Main Mathematics and scored 86% in the January 2025 session and 90.2% in the April session. This was achieved despite it being a relatively small model with 7 billion parameters and trained on a single H100 GPU.
“Sometimes, the right support at the right moment can change a student’s entire path,” said Prateek Maheshwari, co-founder of PW and chair of the Indian Edtech Consortium. “At PW, we believe AI can offer that support, if it’s built with care, context, and purpose.”
The model is powered by 130,000 curated question-answer pairs and uses a combination of supervised fine-tuning (SFT), reinforcement learning (RL), and rejection sampling to improve output quality. PW says this approach grounds the model’s reasoning in pedagogical patterns that align with how students are taught for competitive exams.
Maheshwari noted that Aryabhata 1.0 is part of a larger roadmap. “We’re working to expand Aryabhata 1.0 to cover JEE Advanced and a broader range of mathematical domains.”
PW is inviting educators, developers, and researchers to test the model and provide feedback to help refine future iterations. The goal is to eventually build a suite of subject-centric SLMs that support learning across disciplines in an AI-driven education system.
In other news, Google’s Gemini 2.5 Pro recently scored 336.2 out of 360 marks, outperforming the student who topped the exam with 332 marks in 2025.
The post PhysicsWallah Launches Aryabhata 1.0 AI Model to Help Students with JEE Main Maths appeared first on Analytics India Magazine.
Generated by RSStT. The copyright belongs to the original author.