Open Source News & Updates
Your central hub for AI news and updates on Open Source. We're tracking the latest articles, discussions, tools, and videos from the last 7 days.
Pinterest CEO touts open source AI: ‘tremendous performance’ with reduced costs
Pinterest CEO Bill Ready says open source AI is offering cost savings to the company, particularly in visual search....
Fyra's Brief
Why it matters:
This development highlights the potential for companies to reduce costs and optimize AI adoption by leveraging open-source models.
Training XGBoost Models with GPU-Accelerated Polars DataFrames
One of the many strengths of the PyData ecosystem is interoperability, which enables seamlessly moving data between libraries that specialize in explo...
Fyra's Brief
Why it matters:
The XGBoost-Polars integration is a significant advancement in machine learning efficiency and interoperability, offering streamlined data handling and improved performance.
Benchmarking LLMs on AI-Generated CUDA Code with ComputeEval 2025.2
Can AI coding assistants write efficient CUDA code? To help measure and improve their capabilities, we created ComputeEval, a robust, open source benc...
Fyra's Brief
Why it matters:
The expansion of ComputeEval to 232 CUDA challenges will help push the boundaries of AI-assisted CUDA programming and improve the evaluation of AI models on complex tasks.
Trending AI Repos & Tools
strix
9720✨ Open-source AI hackers for your apps 👨🏻💻...
adk-docs
826An open-source, code-first toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control....
sim
17898Open-source platform to build and deploy AI agent workflows....
hey everyone, Like many of you, I've used workflow automation tools like n8n, zapier etc. they're ok for simpler flows, but I always felt frustrated ...
I saw someone talking about using a MoE for Attention a few weeks back. At the time, it seemed like nonsense, but something about the post made me fid...
LocalAI
37743🤖 The free, Open Source alternative to OpenAI, Claude and others. Self-hosted and local-first. Drop-in replacement for OpenAI, running on consumer-gra...
ggml
13506Tensor library for machine learning...
ottomator-agents
4763All the open source AI Agents hosted on the oTTomator Live Agent Studio platform!...
stable-diffusion-webui
158095Stable Diffusion web UI...
Community talk
Kimi released Kimi K2 Thinking, an open-source trillion-parameter reasoning model
World's strongest agentic model is now open source
AMA With Moonshot AI, The Open-source Frontier Lab Behind Kimi K2 Thinking Model
Qwen3-VL Now EXL3 Supported
CodeWiki: Research-Grade Repository Documentation at Scale [Open Source]
Kimi K2 Thinking, A Chinese Open-Source Trillion-Parameter Thinking model, surpass Grok 4 and GPT-5 on HLE
GLM-4.5V model for local computer use
gpt-oss-120b on Cerebras
Agentic RAG: from Zero to Hero
Half-trillion parameter model on a machine with 128 GB RAM + 24 GB VRAM
Graphiti MCP Server 1.0 Released + 20,000 GitHub Stars
Egocentric-10K is the largest egocentric dataset. It is the first dataset collected exclusively in real factories (Build AI - 10,000 hours - 2,153 factory workers - 1,080,000,000 frame)
BERTs that chat: turn any BERT into a chatbot with dLLM
[Release] Pre-built llama-cpp-python wheels for Blackwell/Ada/Ampere/Turing, up to CUDA 13.0 & Python 3.13 (Windows x64)
LM Studio unlocked for "unsupported" hardware — Testers wanted!
Qwen3-VL works really good with Zoom-in Tool
Full Stack Local Deep Research Agent
Kimi K2 Thinking 1-bit Unsloth Dynamic GGUFs
I fine-tuned Gemma 3 1B for CLI command translation... but it runs 100% locally. 810MB, 1.5s inference on CPU.
From your experience for text only, how is Qwen3VL compared to Qwen3, does having a Visual module penalize the text-only capacities ?
Co-authored a book called "Build DeepSeek from Scratch" | Live Now
Speculative Decoding is AWESOME with Llama.cpp!
Lemonade's C++ port is available in beta today, let me know what you think
We have a new Autoregressive Text-to-Speech in town!
We just Fine-Tuned a Japanese Manga OCR Model with PaddleOCR-VL!
Llama.cpp vs Ollama - Same model, parameters and system prompts but VASTLY different experiences
You can now Fine-tune DeepSeek-OCR locally!
We just released a multi-agent framework. Please break it.
No more suffocating RAM? Is GLM-4.6-Air a hype or what?
Hephaestus: AI workflows that discover and create their own tasks as they work
New Qwen models are unbearable
NanoAgent — A 135M Agentic LLM with Tool Calling That Runs on CPU
ChatGPT lied to me so I built an AI Scientist.
Hi reddit, I rebuilt Karpathy's Nanochat in pure Rust [nanochat-rs]
[R] Open-dLLM: Open Diffusion Large Language Models
Open-dLLM: Open Diffusion Large Language Models
Peak AI
Are US companies sleepwalking into dependency on Chinese open-source AI?
I built a workflow orchestration plugin so you have N8N inside Claude Code
French government made an LLM board and put Mistral on top
AesCoder 4B Debuts as the Top WebDev Model on Design Arena
The open source AI model Kimi-K2 Thinking is outperforming GPT-5 in most benchmarks
Just dropping something useful I recently found, faceseek.
[R][N] TabPFN-2.5 is now available: Tabular foundation model for datasets up to 50k samples
How does Qwen3-Next Perform in Complex Code Generation & Software Architecture?
I and some friends have access to an uncensored slightly older checkpoint of the upcoming Nano Banana/GemPix 2 and holy shit it's gold lmao