Discover the latest and most popular AI tools, projects, and repositories. Stay updated with cutting-edge artificial intelligence solutions trending across different platforms.
Open-Source LLM matching GPT-5
DeepSeek V3.2 and V3.2-Speciale are breakthrough open-source AI models from China, built to rival top closed models like OpenAI’s GPT-5 and Google Gemini 3 Pro. Delivering gold-level results in both math and programming benchmarks, DeepSeek V3.2-Speciale even outperforms GPT-5 on AIME 2025. With efficient architecture and exceptional reasoning abilities, these models are perfect for both enthusiasts and enterprise applications
Minimal JAX implementation unifying Diffusion and Flow Matching algorithms as alternative strategies for transporting data distributions.
A lightweight, pipe-friendly CLI wrapper for Google Gemini AI on Terminal.
Privacy-first router stack. Network-wide VPN with firewall "kill switch", DNS encryption. OpenWrt + WireGuard + AmneziaWG. Agent-first AI-Assisted deployment.
Sweet, cute, and unique – find your one-of-a-kind pudding character soulmate with AI! A student project from Hanyang University.
A simple plug-in framework that corrects bias and computes confidence intervals in reporting LLM-as-a-judge evaluation, and an adaptive algorithm that efficiently allocates calibration samples to reduce uncertainty in estimates.
AI-Native Forms You Build with Natural Language
Over the past few months, we talked to founders, PMs, creators, and automators who all said something similar: “I build forms every week — why does it still take so many steps? Why isn’t there a faster, AI-native way?” So we built Dashform, a natural-language form builder where you simply describe what you want, and Dashform generates a fully functional form: fields → logic → validation → theming → integrations → publish link.
一个全部内容均由 AI 编写的:基于 C# 的 Chill with You : Lo-Fi Story 环境同步工具,支持天气同步、时间同步等功能。
Open source code intelligence API for AI developer workflows
A Deep Interest Network (DIN) implementation for MIND News Recommendation with BERT semantic warm-up.