Today's AI Research: Fastest-Growing Projects — May 11, 2026
Today's AI Research, we're seeing a surge in interest around multimodal intelligence, with several repositories exploring new frontiers in language models, agentic world modeling, and embodied AI. The growth scores are also indicating a strong focus on research applications in finance, healthcare, and game development.
One of the fastest-growing repositories this week is lukiIabs/trading-agents, with a Growth Score of 48.55 and 233 stars. This project provides a multi-agent trading platform for stocks, crypto, and fintech, leveraging OpenAI's LLM capabilities for quantitative algo trading and sentiment analysis. Its rapid growth can be attributed to the increasing interest in AI-powered finance tools.
fkyah3/opencode-yg is another notable repository, with a Growth Score of 16.97 and 36 stars. This research fork demonstrates Language Anchoring, enabling LLMs to think consistently in a specific language, with a verified compliance rate of over 95% for Chinese thinking. Its growth can be attributed to the importance of language understanding in AI research.
matrix-agent/awesome-agentic-world-modeling boasts an impressive 198 stars and a Growth Score of 8.38. This repository provides a comprehensive overview of Agentic World Modeling, covering its foundations, capabilities, laws, and beyond. Its steady growth reflects the ongoing interest in this area of research.
AutoMedBench/AutoMedBench has a Growth Score of 8.03 and 26 stars. As a medical auto-research benchmark for autonomous AI agents, it offers a valuable resource for researchers exploring healthcare applications. The project's growth can be attributed to the increasing focus on AI in medicine.
thunlp/OPD, with a Growth Score of 7.76 and 312 stars, is an official repository for research on On-Policy Distillation of large language models. Its high star count indicates significant interest in this area of study.
XIAO4579/PRISM boasts a Growth Score of 5.53 and 68 stars. This project explores pre-alignment via black-box On-Policy Distillation for multimodal RL, pushing the boundaries of what's possible with SFT-to-RL methods. Its growth can be attributed to the ongoing quest for innovation in reinforcement learning.
gameworld-project/gameworld has a Growth Score of 4.92 and 172 stars. As a standardized evaluation platform for multimodal game agents, it provides a valuable resource for researchers in this area. The project's steady growth reflects its importance in the gaming community.
Hedlen/Awesome-Multimodal-Intelligence is a curated collection of papers, code, and datasets related to VLMs, VLAs, world models, and embodied AI, with a Growth Score of 3.47 and 41 stars. Its growth can be attributed to its value as a resource for researchers exploring these areas.
AMAP-ML/DCW has a Growth Score of 3.40 and 115 stars. This repository explores the SNR-t bias of diffusion probabilistic models, offering insights into the underlying mechanisms. The project's growth reflects interest in this specific area of research.
Lastly, Yovecent/UDM-GRPO boasts a Growth Score of 3.16 and 24 stars. As a stable and efficient group relative policy optimization for uniform discrete diffusion models, it provides an innovative approach to reinforcement learning. Its growth can be attributed to its novelty and potential impact in the field.
One of the fastest-growing repositories this week is lukiIabs/trading-agents, with a Growth Score of 48.55 and 233 stars. This project provides a multi-agent trading platform for stocks, crypto, and fintech, leveraging OpenAI's LLM capabilities for quantitative algo trading and sentiment analysis. Its rapid growth can be attributed to the increasing interest in AI-powered finance tools.
fkyah3/opencode-yg is another notable repository, with a Growth Score of 16.97 and 36 stars. This research fork demonstrates Language Anchoring, enabling LLMs to think consistently in a specific language, with a verified compliance rate of over 95% for Chinese thinking. Its growth can be attributed to the importance of language understanding in AI research.
matrix-agent/awesome-agentic-world-modeling boasts an impressive 198 stars and a Growth Score of 8.38. This repository provides a comprehensive overview of Agentic World Modeling, covering its foundations, capabilities, laws, and beyond. Its steady growth reflects the ongoing interest in this area of research.
AutoMedBench/AutoMedBench has a Growth Score of 8.03 and 26 stars. As a medical auto-research benchmark for autonomous AI agents, it offers a valuable resource for researchers exploring healthcare applications. The project's growth can be attributed to the increasing focus on AI in medicine.
thunlp/OPD, with a Growth Score of 7.76 and 312 stars, is an official repository for research on On-Policy Distillation of large language models. Its high star count indicates significant interest in this area of study.
XIAO4579/PRISM boasts a Growth Score of 5.53 and 68 stars. This project explores pre-alignment via black-box On-Policy Distillation for multimodal RL, pushing the boundaries of what's possible with SFT-to-RL methods. Its growth can be attributed to the ongoing quest for innovation in reinforcement learning.
gameworld-project/gameworld has a Growth Score of 4.92 and 172 stars. As a standardized evaluation platform for multimodal game agents, it provides a valuable resource for researchers in this area. The project's steady growth reflects its importance in the gaming community.
Hedlen/Awesome-Multimodal-Intelligence is a curated collection of papers, code, and datasets related to VLMs, VLAs, world models, and embodied AI, with a Growth Score of 3.47 and 41 stars. Its growth can be attributed to its value as a resource for researchers exploring these areas.
AMAP-ML/DCW has a Growth Score of 3.40 and 115 stars. This repository explores the SNR-t bias of diffusion probabilistic models, offering insights into the underlying mechanisms. The project's growth reflects interest in this specific area of research.
Lastly, Yovecent/UDM-GRPO boasts a Growth Score of 3.16 and 24 stars. As a stable and efficient group relative policy optimization for uniform discrete diffusion models, it provides an innovative approach to reinforcement learning. Its growth can be attributed to its novelty and potential impact in the field.