Super AGI Research Lab

Research lab dedicated to explore and pursue Generalized Super Intelligence

AI Employees Whitepaper

Find our latest research on AI Agents & their impact on businesses

Technical Research Areas

Neurosymbolic AI

Integrating symbolic reasoning with neural networks to achieve more abstract and human-like cognitive abilities.

Autonomous Agents & Multi-Agent Systems

Developing intelligent agents capable of independent decision-making and collaboration within complex environments.

Novel Model Architectures

Exploring innovative architectures and frameworks to enhance the capabilities and efficiency of AI systems.

System 2 Thinking

Investigating higher-order cognitive processes, such as reasoning, planning, and problem-solving, akin to human System 2 thinking.

Recursive Self-Improvement Systems

Designing AI systems capable of autonomously improving their own algorithms, architectures, and learning strategies over time.

Socio-Economic Research Areas

Digital Workforce

Examining the impact of AI and automation on employment dynamics, skill requirements, and the future of work in the digital age.

Algorithmic Governance

Studying the use of AI algorithms in decision-making processes within governance structures, and the associated implications for accountability, transparency, and fairness.

Universal Basic Income (UBI)

Investigating the potential role of UBI in mitigating the socio-economic effects of automation and AI-driven labor market shifts.

Ethical AI

Addressing ethical considerations and challenges in the development, deployment, and governance of advanced AI systems, with a focus on fairness, accountability, transparency, and societal impact.

Human-AI Collaboration

Exploring the dynamics of collaboration between humans and AI systems in various contexts, including work, education, healthcare, and creative endeavors.

Our Publications

V-Zen: Efficient GUI Understanding and Precise Grounding With A Novel Multimodal LLM

GUIDE: Graphical User Interface Data for Execution

AUTONODE: A Neuro-Graphic Self-Learnable Engine for Cognitive GUI Automation

VEagle: Advancements in Multimodal Representation Learning

Recursive Agent Trajectory Fine-Tuning: Utilizing Agent Instructions for Enhanced Autonomy and Efficiency in AI Agents

Research

  • SuperCoder 2.0 achieves 33.66% success rate in SWE-bench Lite, ranking #3 globally & #1 among all open-source coding systems

    Introduction In recent years, there has been a notable surge in the utilization of multi-agent systems powered by state-of-the-art Large Language Models (LLMs) to address a wide array of tasks, [...]

  • Large Coding Models

    The Role of Large Coding Models (LCMs) in Autonomous Software Development

    Introduction In our last blog, we discussed the challenges around Autonomous Software Development which is totally free from human intervention. Toward the end, we shared our stance on why startups [...]

  • SuperAGI - autonomous software development

    Autonomous Software Development is here!

    Introduction The year is 2030. The latest company to get listed on NASDAQ has just 2 employees. There is a CEO, and a CTO and they are supported by a [...]

  • Multi-Agent System

    All of us have heard about the Mixture-of-Experts (MoE) architecture for LLMs. MoE divides models into separate sub-networks (or “experts”), each specializing in a subset of the input data, to [...]

  • Introducing DoRA : The Self Training Module of AutoNode

    Introduction In cognitive process automation, developing self-training modules is crucial. These modules can independently explore, learn, and adapt to complex and unfamiliar environments in the interface. They do this by [...]

  • A Deep Dive into Policy Optimization Algorithms & Frameworks for Model Alignment

    A Comprehensive Exploration of Policy Optimization Algorithms and Frameworks Introduction Reinforcement Learning (RL) is an intriguing area of machine learning that deals with the actions of intelligent agents within an [...]

  • Towards AGI Part 2: Multiverse of Actions

    Towards AGI: [Part 2] Multiverse of Actions

    In part-1 of Towards AGI series, we discussed a core component of Agents - Memory. However, the early agent architectures, didn’t have Memory as a first class primitive. As we [...]

  • Towards AGI: [Part 1] Agents with Memory

    Agents are an emerging class of artificial intelligence (AI) systems that use large language models (LLMs) to interact with the world. In the 'Towards AGI' series, we aim to explore [...]

  • Meet SuperAGI’s VEagle: An Open-source vision model that beats SoTA models like Bliva & Llava

    Introduction VEagle significantly improves the textual understanding & interpretation of images. The unique feature of VEagle is in its architectural change along with a combination of different components: a vision [...]

  • Introducing AutoNode: Advancing RPA with a Multi-Expert AI System

    AutoNode is a significant progression in Robotic Process Automation (RPA), addressing the limitations of current systems through a synergistic integration of specialized AI models. This solution targets the inefficiencies and [...]

  • Small Agentic Model

    Introducing SAM – A 7B Small Agentic Model that outperforms GPT-3.5 and Orca on reasoning benchmarks

    Introduction SuperAGI is focused on developing Large Agentic Models (LAMs) that will power autonomous AI agents. As part of this effort, we have been working on enhancing multi-hop sequential reasoning [...]

  • Introduction to RAGA – Retrieval Augmented Generation and Actions

    Retrieval-augmented generation (RAG) has made significant strides in enhancing the Language Model's (LLM) ability to provide contextual and informed responses by leveraging external knowledge bases. The process chiefly involves an [...]

Spotlight Papers

Research papers we are reading

  • MAGIS: LLM-Based Multi-Agent Framework for GitHub Issue Resolution

  • CodeS: Natural Language to Code Repository via Multi-Layer Sketch

  • DebugBench: Evaluating Debugging Capability of Large Language Models

  • ChatterBox: Multi-round Multimodal Referring and Grounding

  • KOSMOS-2: Grounding Multimodal Large Language Models to the World

  • CogVLM: Visual Expert for Pretrained Language Models