Top 10 Fastest-Growing AI Projects on GitHub in 2026

A curated list of the 10 AI projects with the highest star growth on GitHub in 2026—covering AI agents, multimodal models, and local deployment tools—to help developers stay ahead of emerging trends.

Decision in 20 seconds

A curated list of the 10 AI projects with the highest star growth on GitHub in 2026—covering AI agents, multimodal models, and local deployment tools—to help de…

Who this is for

Founders, Product managers, and Developers who want a repeatable, low-noise way to track AI updates and turn them into decisions.

Key takeaways

• 1. Clawdbot — The First Breakout AI Agent Architecture
• 2. GitHub Agent HQ (OpenAI Codex Integration Edition)
• 3. MiniCPM-o 4.5 — The First Open-Source Full-Duplex Multimodal Model
• 4. Qwen3-Coder-Next

Top 10 Fastest-Growing AI Projects on GitHub in 2026

AI projects on GitHub exploded in growth throughout 2026. As Agentic Engineering became the dominant development paradigm, open-source initiatives increasingly focused on intelligent agents, lightweight deployment, and real-world application. This article highlights the top 10 fastest-growing AI repositories on GitHub as of February 2026—helping developers quickly identify high-potential technical directions.

1. Clawdbot — The First Breakout AI Agent Architecture

Since its launch in early 2026, Clawdbot has amassed over 480,000 stars—making it one of the year’s fastest-growing projects. Its core innovation lies in a modular agent design that supports dynamic task decomposition, tool invocation, and cross-platform collaboration. Built in Rust, it prioritizes low latency and high concurrency—and is already powering internal automation workflows at multiple enterprises. According to RadarAI’s February 4 report, Clawdbot is actively reshaping enterprise-grade agent deployment patterns.

2. GitHub Agent HQ (OpenAI Codex Integration Edition)

A joint initiative by OpenAI and GitHub, Agent HQ deeply integrates the Codex model and is tailored for Copilot Pro+ and Enterprise users. Within its first week, it saw over 500,000 downloads (RadarAI, February 5). It enables automated code generation, test writing, and CI/CD pipeline optimization. Its standout feature is native support for the Model Context Protocol (MCP)—enabling seamless context sharing across platforms.

3. MiniCPM-o 4.5 — The First Open-Source Full-Duplex Multimodal Model

MiniCPM-o 4.5 is the most talked-about lightweight multimodal project of 2026. With just 9 billion parameters, it outperforms GPT-4o on multiple benchmarks. It supports real-time audio-video interaction, proactive notifications, and environmental awareness—making it ideal for edge-device deployment. Its GitHub page shows over 320,000 stars gained in just two weeks, and the community has already spawned numerous mobile and embedded application forks.

4. Qwen3-Coder-Next

Qwen3-Coder-Next, launched by Alibaba’s Tongyi Lab in collaboration with vLLM, topped GitHub Trending on its first day. Optimized specifically for code generation, it supports over 70 programming languages and delivers 3× faster inference than its predecessor. Its deep integration with vLLM significantly lowers the barrier to local deployment—making it the top choice for individual developers building private, on-device coding assistants.

5. Claude Code for Xcode

Anthropic has natively integrated Claude Code into Xcode 26.3, enabling visual verification, cross-project understanding, and autonomous task execution. Though primarily an Apple ecosystem plugin, its open-source core components have already garnered 280K stars on GitHub. This marks a major milestone: agentic programming has officially entered mainstream development environments—especially resonating with iOS and macOS developers.

6. FAY — A Framework for Connecting Digital Humans with LLMs

FAY, originally trending on Gitee and later open-sourced on GitHub, bridges 2.5D/3D digital humans with large language models. It supports APIs from OpenAI, DeepSeek, and other leading providers—and runs seamlessly on web, mobile, and desktop platforms. As of January 2026, it has amassed 210K stars on GitHub and is widely used in virtual customer service, AI-powered tutoring, and similar applications.

7. Bun AI Runtime

Bun—the blazing-fast JavaScript runtime built in Zig—added native AI support in 2026. Its built-in vector computation library and lightweight inference engine let frontend developers run small LLMs directly in browsers or Node.js environments. With star growth topping GitHub’s weekly charts for three consecutive weeks, it’s now recommended as the default runtime by open-source tools like NocoBase (per a blog post on cnblogs.com, February 6, 2026).

8. RAGLite — A Lightweight Retrieval-Augmented Generation Framework

RAGLite tackles the complexity and high resource demands of traditional RAG systems with its “file-as-interface” philosophy—championed by Jerry Liu, co-founder of LlamaIndex. Just one file is enough to index documents and power Q&A. Memory usage stays under 50 MB. It crossed 190K stars on GitHub in just two weeks, quickly becoming a go-to framework for lightweight, local knowledge-base applications.

9. OpenClaw — A Context-Aware Personal AI Assistant

After going viral in 2025, OpenClaw has continued rapid iteration—its 2026 release significantly enhances long-term memory and tool-calling capabilities. Project founder Peter Steinberger describes its mission as “making AI truly remember you.” It now boasts over 1.6 million GitHub Stars, and a vibrant community has produced numerous one-click install scripts and Chinese-language adaptation packages—dramatically lowering the barrier to entry.

10. MCP Apps SDK

With ChatGPT’s full adoption of the MCP standard (announced by RadarAI on February 5), the MCP Apps SDK has emerged as a critical tool for developers integrating cross-platform AI context. This SDK provides a unified interface enabling apps to share user state, conversation history, and tool permissions. It garnered 170,000 GitHub Stars within its first week—widely seen as a pivotal step toward standardization in the AI app ecosystem.

Tool Recommendation: How to Track GitHub AI Project Updates

RadarAI aggregates high-quality AI updates daily—helping developers quickly assess which projects are production-ready. It supports RSS feeds, so you can push alerts directly to Feedly or Inoreader.

Related reading

• Top China-Built AI Models to Watch in 2026: DeepSeek, Qwen, Kimi & More
• China AI Updates in English: What Builders Should Watch Each Month
• How to Track China AI in English Without Doomscrolling
• Best English Sources for China AI Industry Updates (2026 Guide)

RadarAI helps builders track AI updates, compare source-backed signals, and decide which changes are worth acting on.