Advancing Intelligence Frontiers

QuantaAlpha

Elite AI Agent Research Team πŸš€
Building the future of Autonomous Intelligence

Explore Research Collaboration
April 2025
Founded
30+
Elite Researchers
UCAS, PKU, THU, CMU, HKUST, USTC, ZJU.
Top Affiliations

Research Focus

CodeAgent

End-to-end execution for real-world coding tasks, enabling agents to autonomously understand, modify, and run complex repositories.

DeepResearch

Multi-step reasoning agents for open-ended scientific and knowledge discovery, synthesizing insights across domains.

Agentic RL

Reasoning-centered reinforcement learning, bridging decision-making, planning, and adaptive intelligence.

Self-Evolving

Agents that continuously improve through self-evolution and collaborative learning, enabling emergent collective intelligence.

Latest News

2025.11
πŸŽ‰ Check out our latest work: "Octopus: Agentic Multimodal Reasoning with Six-Capability Orchestration" on arXiv!
2025.12
πŸ† Paper "GitTaskBench" accepted to AAAI 2026 as Oral Presentation! (Total 3 papers accepted)
2025.10
πŸ† Paper "RepoMaster" accepted to NeurIPS 2025 as Spotlight! (Total 2 papers accepted)
2025.10
πŸ† Paper "SE-Agent" accepted to NeurIPS 2025 as Poster! (Total 2 papers accepted)
2025.10
πŸŽ‰ Paper "ALRPHFS: Adversarially Learned Risk Patterns..." accepted to Findings of EMNLP 2025.
2025.08
πŸŽ‰ Check out our latest work: "GitTaskBench: A Benchmark for Code Agents Solving Real-World Tasks" on arXiv!
2025.08
πŸŽ‰ Check out our latest work: "CogFSM: Enabling Explicit and Capability-aware Multi-Agent Systems" on arXiv!
2025.08
πŸŽ‰ Check out our latest work: "SE-Agent: Self-Evolution Trajectory Optimization" on arXiv!
2025.05
πŸŽ‰ Paper "Beyond Surface-Level Patterns..." accepted to Findings of ACL 2025.
2025.05
πŸŽ‰ Check out our latest work: "RepoMaster: Autonomous Exploration of GitHub Repositories" on arXiv!
2025.02
πŸ“„ New preprint: "ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs" released on arXiv.

Selected Publications

Octopus

πŸ™ Octopus: Agentic Multimodal Reasoning with Six-Capability Orchestration

QuantaAlpha Team arXiv Preprint

Comprehensive agentic framework orchestrating six distinct capabilities for advanced multimodal reasoning and problem solving.

Paper Code Coming Soon
SE-Agent

🧠 SE-Agent: Self-Evolution Trajectory Optimization in Multi-Step Reasoning

QuantaAlpha Team NeurIPS 2025 Poster

A novel framework for self-evolving agents that optimizes reasoning trajectories through continuous learning and adaptation.

Paper GitHub
GitTaskBench

🧩 GitTaskBench: A Benchmark for Code Agents Solving Real-World Tasks

QuantaAlpha Team AAAI 2026 Oral

A benchmark evaluating code agents on their ability to solve complex, real-world tasks by leveraging entire code repositories.

Paper GitHub
RepoMaster

πŸ” RepoMaster: Autonomous Exploration and Understanding of GitHub Repositories

QuantaAlpha Team NeurIPS 2025 Spotlight

Autonomous agents capable of exploring and understanding large-scale codebases to solve complex programming tasks.

Paper GitHub

Join Our Community

Engage with us on WeChat and XiaoHongShu for real-time discussions

Personal WeChat

Harry_-_Wang
WeChat ID

Add for direct communication with the team core.

Academic Exchange

WeChat QR

Scan to join our vibrant research community group.

XiaoHongShu

QuantaAlpha
εΏ—εŒι“εˆε€„
Click to Copy Link

Follow our updates on RED.

Collaboration & Careers

We are always looking for exceptional talent and partners.

Seeking Partners

  • β€’ University/Corp Labs: Joint research, compute resources.
  • β€’ Startups: Industry applications of Agent tech.
  • β€’ Open Source: Contributors for our GitHub repos.

Ideal Candidate

  • β€’ Top-tier conference publications (First Author).
  • β€’ Strong self-motivation & passion for AGI.
  • β€’ Focus: Agentic RL, CodeAgents, Reasoning.
quantaalpha.ai@gmail.com

Reach out to explore unlimited possibilities.