We present AlphaApollo, an agentic reasoning system that targets two bottlenecks in foundation-model reasoning: (1) limited reasoning capacity for complex, long-horizon problem solving and (2) unreliable test-time evolution without trustworthy verification. AlphaApollo orchestrates models and tools via three components: (i) multi-turn agentic reasoning, which formalizes model-environment interaction with structured tool calls and responses; (ii) multi-turn agentic learning, which applies turn-level reinforcement learning to optimize tool-use reasoning while decoupling actions from tool responses for stable training; and (iii) multi-round agentic evolution, which refines solutions through a propose-judge-update loop with tool-assisted verifications and long-horizon memory. Across seven math reasoning benchmarks and multiple model scales, AlphaApollo improves performance through reliable tool use (> 85% tool-call success), substantial gains from multi-turn RL (Avg@32: Qwen2.5-1.5B-Instruct 1.07% -> 9.64%, Qwen2.5-7B-Instruct 8.77% -> 20.35%), and improvements from evolution (e.g., Qwen2.5-3B-Instruct 5.27% -> 7.70%, Qwen2.5-14B-Instruct 16.53% -> 21.08%). This project is still ongoing. We welcome feedback from the community and will frequently update the source code and technical report.
@article{arxiv.2510.06261,
title = {AlphaApollo: A System for Deep Agentic Reasoning},
author = {Zhanke Zhou and Chentao Cao and Xiao Feng and Xuan Li and Zongze Li and Xiangyu Lu and Jiangchao Yao and Weikai Huang and Tian Cheng and Jianghangfan Zhang and Tangyu Jiang and Linrui Xu and Yiming Zheng and Brando Miranda and Tongliang Liu and Sanmi Koyejo and Masashi Sugiyama and Bo Han},
journal= {arXiv preprint arXiv:2510.06261},
year = {2026}
}