Open-Source AI
Hugging Face, GitHub, Ollama, local models, agent frameworks
The lead
Latest first
Latest in Open-Source AI
llama.cpp Creator Uses Qwen3-27B for 6 Weeks of Local Coding
ggml.org maintainer says Qwen3-27B is his daily driver on M2 Ultra and RTX 5090, highlighting a split between local-first and cloud-integrated AI…
Allen AI releases MolmoMotion for language-guided 3D motion
Allen AI open-sources MolmoMotion, a model that predicts 3D object trajectories from language instructions, plus the 1.16M-video MolmoMotion-1M dataset and…
Local AI Models Hit Viability for Daily Development
Local AI Models Hit Viability for Daily Development: Vicki Boykis reports Gemma 4 and GPT-OSS enable agentic coding at 75% of frontier…
GitHub Releases Open Multilingual Dataset for AI Developers
GitHub Releases Open Multilingual Dataset for AI Developers: GitHub publishes 40M-repo metadata dataset under CC0-1.0, revealing Portuguese leads non-English…
How to Evaluate LLMs with olmo-eval (AllenAI, 2026)
Step-by-step guide to olmo-eval — AllenAI's lightweight evaluation framework that runs benchmarks directly (no containers by default), supports per-question…
Qwen3 flagship open model rivals DeepSeek-R1, OpenAI o1
Alibaba's Qwen3 (April 2025) drops 8 models (0.6B–235B), hybrid thinking modes, 119 languages, Apache 2.0 license. Benchmarks vs DeepSeek-R1, o1, coding…
Open Source AI Spring 2026: HF + GitHub Growth Data
Hugging Face Spring 2026 report: 13M users, 2M+ models, China leads downloads, 39% independent devs. What it means for your stack.
WASM Browser Local AI 2026: What Developers Actually Ship
Real projects using Pyodide, WebLLM, Transformers.js for WASM browser local AI inference. Verified examples from GitHub, HN, production apps — no demos.
DeepSeek V4-Pro’s MIT License Changes Everything
DeepSeek V4-Pro ships MIT-licensed with a 1M-token context and 93.5 LiveCodeBench — the open model that finally closes the legal gap with…
Run a Local LLM With Ollama in 5 Minutes
How to run a local LLM with Ollama in 2026: install, pick a model sized for your RAM, start chatting, hit the…