Radar

Open source

Free AI models and tools that anyone can download. Newest first.

Importance 2 of 5Hacker News (AI, 100+ points)

Essay argues all AI-generated code should be open source

A widely discussed blog post argues that code generated by AI should be released as open source. It gathered nearly 900 points on Hacker News.

Why it matters: The debate touches on licensing and ownership questions companies face when shipping AI-written code.

Importance 2 of 5r/LocalLLaMA

Meituan releases open-source LongCat-Video-Avatar 1.6 for audio-driven video

Meituan released LongCat-Video-Avatar 1.6 on Hugging Face, an upgrade to its open-source framework for audio-driven human video generation. It improves lip sync, facial and body motion, and long-form consistency, and adds inference optimizations.

Why it matters: Open avatar video models lower the cost of building synthetic presenters and video content in-house.

Importance 4 of 5The Decoder

Meituan releases open weights for LongCat-2.5, a 1.6T-parameter agent model trained on Chinese chips

Meituan released the weights for LongCat-2.5, a 1.6-trillion-parameter model built for AI agents and trained entirely on Chinese hardware. It scores 88.4 on Terminal Bench, just behind Claude Opus 5. The sample deployment runs on 194 Huawei Ascend chips with no mention of Nvidia GPUs.

Why it matters: A near-frontier open agent model trained without Nvidia hardware shows Chinese labs can work around export controls, widening open-weight options for enterprises.

Only want news that affects your AI projects?
Track them in Radar and get an alert when a vendor changes something. Free to start.
Start free
Importance 2 of 5Simon Willison 2 sources

Datasette Agent 0.6a0 and Datasette Apps 0.4a0 add homepage agent prompt and app linking

Simon Willison released datasette-agent 0.6a0, which adds an agent prompt to the Datasette homepage with hash-mentions of tables and queries. datasette-apps 0.4a0 adds URL fragment linking between sandboxed apps, and reports link schemes to the agent so it can help write linked apps.

Why it matters: Shows a lightweight open-source pattern for letting AI agents query and build small apps over internal data.

Importance 2 of 5r/MachineLearning

ACM Multimedia 2026: TimeCues Studio - a free, open-source tool for annotating music structure [R]

Hey all, I wanted to share TimeCues Studio with you. It's part of my PhD work (ACM Multimedia 2026), and I thought you might find it useful. It's an open-source web app for annotating music on a beat-locked timeline and comparing algorithms against your annotations. If you need to annotate songs (sections, cues,…

Importance 2 of 5r/LocalLLaMA

LFM2.5-24B-A2B + Hermes is awesome!

I just wanted to share some love for a model that doesn't seem to get very much: LFM2.5-24B-A2B! ( https://huggingface.co/LiquidAI/LFM2.5-24B-A2B-GGUF ) LFM2.5-24B-A2B seems to be the only model that strikes a great balance between speed and capability on commodity hardware. Just to get it out of the way: This is NOT…

Importance 2 of 5r/LocalLLaMA

Open-source Leech lattice quantization brings Qwen3 models to about 2.7 bits per weight

An independent developer released a Rust and CUDA implementation of Qualcomm's LLVQ Leech lattice quantization method and published quantized Qwen3 4B, 8B and 14B models at roughly 2.7 bits per parameter. The developer reports better quality than IQ2_S at the same size but slower inference.

Why it matters: Better low-bit quantization could let teams run capable models on cheaper local hardware, though speed tradeoffs remain.

Importance 2 of 5Google Developers Blog

A2UI 1.0: Build delightful, fast, reliable, agentic experiences

A2UI 1.0 is an open protocol for generative UI that enables AI agents to render interactive, native screens across web, mobile, and desktop applications. Developers can leverage open-source renderer SDKs—including React, Angular, Lit, Flutter, SwiftUI, and Jetpack Compose—alongside agent SDKs for Python, TypeScript,…

Importance 2 of 5Google Developers Blog

Integrating TorchTPU into TorchTitan and Optimizing Models for TPU

TorchTPU integrates a native torch.device("tpu") backend directly into PyTorch, allowing developers to train large-scale models on Google Cloud TPUs without rewriting code in JAX. The TorchTitan TPU fork enhances this with production-grade scaling primitives like SimpleFSDP, Graph Trainer, and custom Pallas kernels…

Importance 2 of 5r/LocalLLaMA

Update: tandem.cpp — DeepSeek-V4.1 on a 64 GB PC with a 16 GB card, and AMD (ROCm) support

Quick update on tandem.cpp, my patch set for llama.cpp that runs MoE models too big for your VRAM faster. It keeps the hot experts on the GPU and lets the CPU help with the rest. Repo: https://github.com/ewhacc/tandem.cpp DeepSeek-V4.1-Flash IQ2_XS (ISTA GSQ-RCO, 139 GiB of experts) on normal PCs Decode tok/s, one…

Importance 2 of 5r/LocalLLaMA

nex-agi/Nex-N2.6 · Hugging Face

Nex-N2.6 is a multimodal agentic model from Nex-AGI, built on GLM-5.3-Flash . It is designed for computer use, coding, and long-horizon tasks involving tools. This repository contains the FP8-quantized weights of Nex-N2.6, with approximately 321B parameters. Following Nex-N2.5-Pro , this release focuses on reducing…

Importance 2 of 5r/LocalLLaMA

Qwen4 Flash wishlist

After using 3.8QFN for some time I have to the conclusion that my wish list for QF4 now only consists of these points. The model can talk to you in a human understandable language See point 1 General improvement in reasoning/intelligence, but mostly see point 1 That's about it. If I could graft 3.8 27B…