Aryabhata: An exam-focused language model for JEE Math
On the map: LLM reasoning → Mathematical reasoning
Abstract (arXiv)
We present Aryabhata 1.0, a compact 7B parameter math reasoning model optimized for the Indian academic exam, the Joint Entrance Examination (JEE). Despite rapid progress in large language models (LLMs), current models often remain unsuitable for educational use. Aryabhata 1.0 is built by merging strong open-weight reasoning models, followed by supervised fine-tuning (SFT) with curriculum learning on verified chain-of-thought (CoT) traces curated through best-of-n rejection sampling. To further boost performance, we apply reinforcement learning with verifiable rewards (RLVR) using A2C objective with group-relative advantage estimation along with novel exploration strategies such as Adaptive Group Resizing and Temperature Scaling. Evaluated on both in-distribution (JEE Main 2025) and out-of-distribution (MATH, GSM8K) benchmarks, Aryabhata outperforms existing models in accuracy and efficiency, while offering pedagogically useful step-by-step reasoning. We release Aryabhata as a foundation model to advance exam-centric, open-source small language models. This marks our first open release for community feedback (https://huggingface.co/PhysicsWallahAI/Aryabhata-1.0); PW is actively training future models to further improve learning outcomes for students.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Mathematical reasoning, one of 17 directions in LLM reasoning. Builds on 15 of its 19 references that are in the map. First posted on arXiv 12 Aug 2025.
Builds on
15 of 19 references.
Closest ancestors · 2
Same direction, newest first
- 2024AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modeling19 Dec
- 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models5 Feb
Other references · 13
- 2014Adam: A Method for Stochastic Optimizationcited by 37,240
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
- 2021Training Verifiers to Solve Math Word Problemscited by 10,109
- 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
- 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
- 2016Asynchronous Methods for Deep Reinforcement Learningcited by 2,258
- 2024HybridFlow: A Flexible and Efficient RLHF Frameworkcited by 2,057
- 2022Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference timecited by 1,396
- 2024Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvementcited by 1,090
- 2024Arcee's MergeKit: A Toolkit for Merging Large Language Modelscited by 194
- 2025AceReason-Nemotron 1.1: Advancing Math and Code Reasoning through SFT and RL Synergycited by 77
- 2025Unlocking Efficient Long-to-Short LLM Reasoning with Model Mergingcited by 58
1 more
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Aryabhata 2: Scaling Reinforcement Learning for Advanced STEM Reasoning
- 2025JT-Math: A Multi-Stage Framework for Advanced Mathematical Reasoning in Large Language Models
- 2026IRIS: Interleaved Reinforcement with Incremental Staged Curriculum for Cross-Lingual Mathematical Reasoning
- 2024OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Data
- 2026GanitLLM: Difficulty-Aware Bengali Mathematical Reasoning through Curriculum-GRPO
- 2026MathShikkha: A Controlled Study of Answer-Only and Chain-of-Thought Supervision for Bangla Mathematical Reasoning in Small Language Models
- 2023WizardMath: Empowering Mathematical Reasoning for Large Language Models via Reinforced Evol-Instruct
- 2024Benchmarking Large Language Models for Math Reasoning Tasks