It’s Next.js Liberation Day. The #1 request we kept hearing: help us run Next fast and secure, without the lock-in and the costs. So we did it. We kept the amazing DX of Next.js, without the bespoke tooling, built on @vite. We’re working with other providers to make deployment

ByteDance just published something I've been waiting for someone to build: CUDA Agent! It trained a model that writes fast CUDA kernels. Not just correct ones — actually optimized ones. It beats torch.compile by 2× on simple/medium kernels, ~92% on complex ones, and even

LTX-2.3 is a major upgrade. It’s a production-ready multimodal engine - designed to be built on. Here’s what’s new 🧵 1/7

LTX-2.3 GGUFs! Quantized models by Unsloth! dev/distilled Q2_K-Q8_0 + text encoders and audio/video vae huggingface.co/unsloth/LTX-2.…

Learn how to run Qwen3.5 locally using Claude Code. Our guide shows you how to run Qwen3.5 on your server for local agentic coding. We then build a Qwen 3.5 agent that autonomously fine-tunes models using Unsloth. Works on 24GB RAM or less. Guide: unsloth.ai/docs/basics/cl…

Today we're releasing our first open source TTS model, TADA! TADA (Text Audio Dual Alignment) is a speech-language model that generates text and audio in one synchronized stream to reduce token-level hallucinations and improve latency. This means: → Zero content hallucinations

i open-sourced autokernel -- autoresearch for GPU kernels you give it any pytorch model. it profiles the model, finds the bottleneck kernels, writes triton replacements, and runs experiments overnight. edit one file, benchmark, keep or revert, repeat forever. same loop as

Anyone who is interested in working at a frontier lab must read this tech report from Nvidia The data engineering section is amazing and look at the amount of different models they used for synthetic data gen research.nvidia.com/labs/nemotron/…

Introducing Mixedbread Wholembed v3, our new SOTA retrieval model across all modalities and 100+ languages. Wholembed v3 brings best-in-class search to text, audio, images, PDFs, videos... You can now get the best retrieval performance on your data, no matter its format.

🚀 Introducing Qianfan-OCR: a 4B-parameter end-to-end model for document intelligence. One model. No pipeline. Table extraction, formula recognition, chart understanding, and key information extraction, all in a single pass. Paper: arxiv.org/abs/2603.13398 Models:

Are you into vibe coding? ✨✨ We have great news for you!! We created a series of GSAP Skills that you can use with your agent of choice 🤖🤖 (Cursor, Claude, Codex, Windsurf , Copilot, 40+ agents). Grab them here and start using them 🦾🦾 github.com/greensock/gsap…

Seedance 2.0 is impressive. But it's closed-source! Introducing our daVinci-MagiHuman — a single-stream 15B Transformer trained from scratch that jointly generates video + audio. No cross-attention. No multi-stream branches. Just self-attention. ⚡ 5s 1080p video in 38s on a

Today we're releasing MolmoWeb, an open source agent that can navigate + complete tasks in a browser on your behalf. Built on Molmo 2 in 4B & 8B sizes, it sets a new open-weight SOTA across four major web-agent benchmarks & even surpasses agents built on proprietary models. 🧵

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

> train embeddings model on actual web search > use it in actual production (200M daily queries) > see crazy results: best contextual retrieval in the world (81.96% CoNTEB; next closest 79.45%) > open source it > 1M hugging face downloads in ~2 weeks > 5-30x cheaper than existing

Covo-Audio (7B) -full-duplex LALM from Tencent. - Qwen2.5-7B + Whisper - Listens and speaks simultaneously (barge-in support). - No separate ASR or TTS pipelines. - Decoupled intelligence/speaker for voice cloning. - 8M hours of audio training. huggingface.co/tencent/Covo-A…

🔊Introducing Voxtral TTS: our new frontier open-weight model for natural, expressive, and ultra-fast text-to-speech 🎭Realistic, emotionally expressive speech. 🌍Supports 9 languages and accurately captures diverse dialects. ⚡Very low latency for time-to-first-audio. 🔄Easily

Blogpost: mistral.ai/news/voxtral-t… Playground: console.mistral.ai/build/audio/te… Technical report: mistral.ai/static/researc… Model weights: huggingface.co/mistralai/Voxt…

Cool. A local, free Topaz+NanoBanana alternative, almost StepFun just dropped a real gift: RealRestorer, a fully open-source image restoration model. - removes rain, fog, glare - no more moiré, JPEG artifacts - restore low-light - 1024×1024 based on Step1X-Edit DiT + Flux-VAE

We build actual open source including dataset Multilingual TTS more than 150 languages with Voice Cloning. It outperforms Dia-TTS, Orpheus, Qwen3-TTS, Chatterbox, and Fish Audio S2 Pro based on automatic CER and MOS!