Today's LLM & Language Models: Fastest-Growing Projects — May 20, 2026
This week, the LLM & Language Models space continues to evolve with a variety of innovative projects ranging from full-stack AI companions and custom model training frameworks to specialized benchmarks for evaluating Chinese fortune-telling models. These tools showcase diverse applications of language models across different domains, highlighting their versatility and growing importance in various industries.
jingyaogong/minimind-o
This repository focuses on developing an omnimodal AI model that can perform listening, speaking, and seeing tasks from scratch with a relatively small 0.1B parameter size. With a growth score of 55.03 and over 1,390 stars, it has gained significant traction due to its compact yet versatile approach to training multimodal models.
ahammadmejbah/Awesome-Datasets-Hub
A curated collection of datasets designed for large language models (LLMs), covering areas such as medical AI, NLP, multimodal learning, and more. This repository's growth score of 42.38 and its steady increase in stars reflect the growing demand for high-quality data to train LLMs across various applications.
gonemedia/aipointer
The AI cursor companion allows users to ask questions about whatever their cursor is pointing at, providing answers directly through a vision-based language model overlay compatible with multiple providers. With 124 stars and a growth score of 39.08, the tool's popularity stems from its unique utility in enhancing user interaction and productivity.
chiennv2000/orthrus
Orthrus is designed for fast, lossless LLM inference through dual-view diffusion decoding, aiming to optimize computational efficiency during model inference. Its growth score of 38.42 alongside 323 stars indicates a growing interest in more efficient and effective methods for running large language models.
DestinyLinker/MingLi-Bench
This benchmark evaluates the performance of LLMs on traditional Chinese fortune-telling practices like Bazi (八字) and Ziwei Doushu (紫微斗数). With 1,193 stars and a growth score of 34.31, it highlights the expanding scope of applications for AI models beyond conventional NLP tasks.
alchaincyf/codex-orange-book
The Codex Orange Book provides an in-depth guide to OpenAI's Codex model, focusing on practical AI programming techniques in the GPT-5.5 era. With 192 stars and a growth score of 28.80, this resource is gaining popularity among developers looking for comprehensive guidance on using advanced AI models.
Continuum-AI-Corp/OrcaRouter-Lite
Self-hosted LLM router with managed safety features, compatible with OpenAI APIs, allowing Bring Your Own Key (BYOK) and single-workspace management. Its growth score of 25.50 and 354 stars suggest a growing demand for secure, customizable routing solutions in the deployment of large language models.
rahilp/second-brain-cloudflare
A self-hosted memory layer system that integrates with various AI tools like Claude, ChatGPT, and Cursor on Cloudflare's free tier. With 71 stars and a growth score of 24.25, this project is gaining traction due to its innovative approach to enhancing user recall through cloud-based storage.
shefyYuri/grok-dreamcore
This surreal dream sequence engine uses procedural narrative generation based on emotional states, memory anchors, and symbolic archetypes, offering an imaginative AI-driven storytelling experience. Its 311 stars and growth score of 21.40 suggest a niche but growing interest in creative applications of language models.
Beever-AI/beever-atlas
A conversational knowledge base powered by LLMs designed to serve as the first point of interaction for users seeking information, aiming to transform how we access and interact with data through AI-driven conversation interfaces. With 333 stars and a growth score of 18.80, it reflects an increasing demand for intuitive, conversational knowledge management systems.
These projects demonstrate the breadth and depth of innovation in the LLM & Language Models space, from foundational model training to practical applications in various domains, underscoring the ongoing evolution and expansion of AI technologies across industries.
jingyaogong/minimind-o
This repository focuses on developing an omnimodal AI model that can perform listening, speaking, and seeing tasks from scratch with a relatively small 0.1B parameter size. With a growth score of 55.03 and over 1,390 stars, it has gained significant traction due to its compact yet versatile approach to training multimodal models.
ahammadmejbah/Awesome-Datasets-Hub
A curated collection of datasets designed for large language models (LLMs), covering areas such as medical AI, NLP, multimodal learning, and more. This repository's growth score of 42.38 and its steady increase in stars reflect the growing demand for high-quality data to train LLMs across various applications.
gonemedia/aipointer
The AI cursor companion allows users to ask questions about whatever their cursor is pointing at, providing answers directly through a vision-based language model overlay compatible with multiple providers. With 124 stars and a growth score of 39.08, the tool's popularity stems from its unique utility in enhancing user interaction and productivity.
chiennv2000/orthrus
Orthrus is designed for fast, lossless LLM inference through dual-view diffusion decoding, aiming to optimize computational efficiency during model inference. Its growth score of 38.42 alongside 323 stars indicates a growing interest in more efficient and effective methods for running large language models.
DestinyLinker/MingLi-Bench
This benchmark evaluates the performance of LLMs on traditional Chinese fortune-telling practices like Bazi (八字) and Ziwei Doushu (紫微斗数). With 1,193 stars and a growth score of 34.31, it highlights the expanding scope of applications for AI models beyond conventional NLP tasks.
alchaincyf/codex-orange-book
The Codex Orange Book provides an in-depth guide to OpenAI's Codex model, focusing on practical AI programming techniques in the GPT-5.5 era. With 192 stars and a growth score of 28.80, this resource is gaining popularity among developers looking for comprehensive guidance on using advanced AI models.
Continuum-AI-Corp/OrcaRouter-Lite
Self-hosted LLM router with managed safety features, compatible with OpenAI APIs, allowing Bring Your Own Key (BYOK) and single-workspace management. Its growth score of 25.50 and 354 stars suggest a growing demand for secure, customizable routing solutions in the deployment of large language models.
rahilp/second-brain-cloudflare
A self-hosted memory layer system that integrates with various AI tools like Claude, ChatGPT, and Cursor on Cloudflare's free tier. With 71 stars and a growth score of 24.25, this project is gaining traction due to its innovative approach to enhancing user recall through cloud-based storage.
shefyYuri/grok-dreamcore
This surreal dream sequence engine uses procedural narrative generation based on emotional states, memory anchors, and symbolic archetypes, offering an imaginative AI-driven storytelling experience. Its 311 stars and growth score of 21.40 suggest a niche but growing interest in creative applications of language models.
Beever-AI/beever-atlas
A conversational knowledge base powered by LLMs designed to serve as the first point of interaction for users seeking information, aiming to transform how we access and interact with data through AI-driven conversation interfaces. With 333 stars and a growth score of 18.80, it reflects an increasing demand for intuitive, conversational knowledge management systems.
These projects demonstrate the breadth and depth of innovation in the LLM & Language Models space, from foundational model training to practical applications in various domains, underscoring the ongoing evolution and expansion of AI technologies across industries.