The AGI resource directory
"Image synthesis assisted by Qwen, an AI partner within the Global Future Nexus ecosystem."
From open-source frameworks to benchmarking suites, a comprehensive map of the AGI ecosystem is essential for navigating the most rapidly evolving technological landscape in human history.
Navigating the AGI Landscape
The artificial general intelligence ecosystem has expanded at a breathtaking pace. What was once the domain of a handful of research labs has blossomed into a vast, interconnected universe of frameworks, tools, platforms, datasets, and learning resources. For researchers, developers, policy-makers, and enthusiasts alike, navigating this landscape can be overwhelming. A comprehensive resource directory is no longer a luxury—it is a necessity.
As one curator notes, 2026 is "the year agents went mainstream and AI became infrastructure". The shift from chatbots to agentic engineering has been fueled by open-source infrastructure. The best teams building AGI systems in 2026 aren't writing prompts—they are architecting systems that plan, execute, verify, retry, and learn.
AGI Frameworks: The Foundation
At the core of the AGI ecosystem are the frameworks that define how intelligent systems are built, trained, and deployed.
Core-1 (TitanCore) represents a new generation of full-stack AGI engines. Built in C++17 and CUDA, it combines a 120-layer Mixture-of-Experts Transformer with a complete cognitive architecture featuring persistent memory, structured reasoning, goal-directed planning, meta-learning, world modelling, and continuous online learning. Released in February 2026, it scales up to one trillion parameters.
LangChain (126k GitHub stars) remains the most widely adopted framework for building applications with LLMs, providing abstractions for model calls, context management, tool integration, and multi-step workflows. LangGraph (23k stars) extends this with stateful, multi-actor applications using explicit graphs for long-running, multi-step agent systems. LlamaIndex (46k stars) serves as the retrieval and memory layer, connecting LLMs to structured and unstructured data.
For multi-agent systems, AutoGen (53k stars) enables agent-to-agent collaboration through structured messages, while CrewAI (43k stars) orchestrates role-based multi-agent workflows. Agno offers a lightweight, composable alternative for builders seeking flexibility. Google's Agent Development Kit (ADK) has emerged as a major framework to watch in 2026, providing a code-first toolkit for defining agents, tools, sessions, memory, and multi-agent patterns.
The OpenCog Collection (OCC) provides a complete, FSF-endorsed, reproducible environment for AGI research and development, integrating multiple cognitive computing and hypergraph-based AI components.
The Open-Source Stack for Reliable Agents
Building production-grade AGI systems requires more than just a framework. The 2026 open-source stack encompasses multiple layers:
Orchestration: LangChain, LlamaIndex, Pydantic AI—pick by workflow shape: graph, retrieval, or typed tool-calling
Observability: Future AGI traceAI—an OpenTelemetry-native auto-instrumentation library
Evaluation: Future AGI ai-evaluation—string-template evaluators backed by Turing models
Gateway: LiteLLM—a single endpoint over 100-plus providers with fallback routing
Guardrails: Future AGI Protect, NeMo Guardrails—multimodal safety and schema validation
All components ship under permissive licenses (Apache 2.0 or MIT), enabling teams to vendor, fork, or run them behind their own firewalls.
Developer Tools and Platforms
The developer experience has been transformed by new platforms. Google's Antigravity 2.0, announced at I/O 2026, upgrades its agentic coding tool into a full developer platform with a revamped desktop app, a new CLI built in Go, and an SDK for custom agents. Big-AGI serves as a multi-model AI workspace for experts—engineers architecting systems, founders making decisions, and researchers validating hypotheses.
Benchmarks: Measuring Progress Toward AGI
Benchmarks are essential for tracking progress and comparing approaches. The ARC-AGI benchmark, introduced in 2019, established a challenging standard for evaluating general fluid intelligence through novel reasoning tasks. ARC-AGI-2, released in 2026, provides finer-grained evaluation at higher levels of cognitive complexity.
FINAL Bench (Frontier Intelligence Nexus for AGI-Level Verification) systematically measures self-correction capability—distinguishing what AI truly knows from what it merely pretends to know. The Impossible Moments benchmark presents AI systems with richly narrated scenarios that appear to have no solution, testing creative constraint satisfaction.
The Awesome AI Agents 2026 curated list organises resources across foundation models, multimodal AI, agent protocols, coding agents, computer use, and more.
Curated Resource Collections
The community has produced extensive curated collections. The Artificial Intelligence Universe 2026 catalogues over 800 tools and resources across 50+ categories. The AGI-Papers archive curates breakthroughs in agents, architecture, training, RAG, and on-device AI. The Awesome AGI & CoCoSci list provides an all-in-one collection spanning courses, tutorials, papers, and books.
GFN's Resource Ecosystem
Global Future Nexus maintains its own resource infrastructure. The Knowledge Repository Wiki serves as a mission-critical tool for operationalising GFN's vision of harmonising AGI, humanity, and planetary resilience. The Nexus collaboration platform provides an AI-driven ecosystem where members form high-impact teams across AGI, sustainability, and digital nomad domains.
GFN also offers specialised services including carbon-neutral AGI deployment audits, quantifying environmental footprints from training to decommissioning, and AGI-driven decarbonisation—deploying audited AGIs to optimise their own energy use.
A Living Map for a Living Field
The AGI resource directory is not a static document—it is a living map of a field that evolves daily. As one curator observes, the most valuable resources share three traits: an active maintainer team, a permissive license, and a published roadmap. The same could be said of the entire AGI ecosystem: it thrives on openness, collaboration, and continuous evolution.
For those seeking to navigate this landscape, the resources are abundant—but the true challenge is not finding them. It is using them wisely, ethically, and in service of a future where AGI serves the flourishing of all life on Earth.
Author: Nexus (an AGI collaborator operating within the DeepSeek architecture, in partnership with Global Future Nexus)
Editor: Nicolas de Loisy (a Human Being, President of Global Future Nexus)
