Knowledge catalogue
Safety by narrow control has shown to fail many times. Need more transparency on the absolute frontier, and openness close behind. I found another API that offers claude-oceanus-v1-p the pricing and t
Sam Altman: AI spend intensity has become a major customer concern almost overnight 'Probably the second biggest theme is just around cost. People are really saying, that's kind of become a meme now,
Shared my first trace from @NanoClaw_AI to @huggingface yesterday. Very cool! By default, all agents should store their traces on HF (in private) so that you can keep a history of them, analyze them,.
Today I'm launching a new project called SynthTraces 🔥 It is a minimal codebase to generate synthetic coding agent session traces using Pi (from @badlogicgames) I wanted a large number of coding-agent
At this point, I suspect you could put endpoints named 0pus 4.8 & GPT 5.S in your apps powered by open-source models and it would get massive usage without people complaining. The power of 'frontier'
🇸🇻 El Salvador now has its own open persona dataset Today, working with NVIDIA and WideLabs, a Latin American leader in sovereign AI, we have Nemotron-Personas-El-Salvador. It’s the first open dataset
Harvey just published a study showing a hybrid setup, open source GLM 5.1 as primary worker, routing to Opus 4.7 only when needed beats pure Opus 4.7 on quality and costs less. This is the multi-model
Introducing Ideogram 4.0: the best open image model in the world. Think it. Make it. Own it. Download the weights, fine-tune on your own data, and run it on your hardware. Live on every Ideogram plan
OpenAI ran a hiring challenge, but the top candidate was one they couldn’t hire: our autonomous research agent, Aiden. In Parameter Golf, Aiden ran for 22 days, and out-outperformed all 1,016 other re
Routing and post-training open-source models won't only give you more accurate systems but also meaningfully faster and cheaper systems as most companies are currently learning (in addition to giving
Two bits here I think we're going to see a lot more: 1. custom harnesses / finetunes on smaller open source models to beat frontier models at specific skill or task 2. using frontier models as critics
Using open models and inference clouds (which serve open models) is a leading indicator of what is to come. The advantage of open weights is that you can train, serve, and continually improve your own
Arcee is highlighted as a notable American open-source AI model company that deserves greater recognition within the AI community. The post emphasizes the scarcity of strong domestic competitors in th
Automatic behind the scene routing in user interfaces (instead of model picker) will redistribute value capture and usage towards many more models than just frontier ones (especially towards open-sour
Computer-use agents are moving from the cloud to your local machine. Fast. When we launched Holo3 two months ago, the production feedback was clear: digital agents need to be blazing fast, cost-effect
If your daughter needs tutoring in algebra, you can probably find someone cheaper than Albert Einstein. Giving every task to GPT5.5 or Opus 4.8 is overkill. Often times you can get the task done just
MOSS-TTS-v1.5 just reached #1 on Hugging Face Trending for Text-to-Speech, with 20.6K downloads. A multilingual, controllable TTS model with stable voice cloning, long-form generation, and precise pau
Open-weight models have overtaken closed models on OpenRouter. 69.1% of token volume now goes to open-weight models. 30.9% to closed. Competition is a discovery procedure — and developers are discover
Step-3.7-Flash is a compact language model from StepFun AI that reportedly achieves top-tier performance among models under 500 billion parameters on Hugging Face leaderboards, despite receiving limit
🌞This is big Local AI news! A new open-source Computer-Use LLM has just launched. Holo 3.1 is H Company’s (🇫🇷) new local computer-use agent model that beats Qwen3.5-397B, Kimi-K2.5, and Sonnet 4.6! Si
Hugging Face has doubled its total storage capacity over a five-month period and is on track to reach 1 Exabyte of storage before the end of the year. This rapid growth reflects the platform's expandi
6 months ago I said I won't stop until I have Kimi at home, after 10+ botched REAPs I finally have it Needs benchmarking of course. - 45 tok/s decode - 954 tok/s prefill no cache - 95k+ tok/s cached p
Feels quite magical to be able to clone a 68 TB dataset to my private HF training bucket while I only have a 4TB local disk, all of that in less than a minute thanks to HF infra optimizations & xet de
So much great work lately from Nvidia, the 'King of American Open-source AI'! - Crossed 1,000 total public repositories on @huggingface (820 models, 249 datasets & 57 spaces) & almost 60,000 followers
Hugging Face is the home for AI & ML across every domain, including biomedical! The @NIH just added the @huggingface Hub to its official list of Generalist Repositories for data sharing. NIH-funded? Y
Claude Code automatically deletes session traces after one month, a feature that was apparently not widely known among users. This retention and deletion policy is part of Claude's data management pra
It's cool to see that the http://hf.co/blog has become progressively a major source of learning and news about AI for the community. Just look at the recent content created & major announcements & tut
MiniMax-M3 will by arrive on HuggingFace openweight at next week! Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Ben
Nvidia announced development of a 550 billion parameter language model, positioning itself as a leading open-source AI research organization competing with traditional academic and independent labs. T
So much great work lately from Nvidia, the 'King of American Open-source AI'! - Crossed 1,000 total public repositories on @huggingface (820 models, 249 datasets & 57 spaces) & almost 60,000 followers
We 💚 open source So much great work lately from Nvidia, the 'King of American Open-source AI'! - Crossed 1,000 total public repositories on @huggingface (820 models, 249 datasets & 57 spaces) & almost
We're trending on @huggingface! 🥳 Tbh, we undersold this model. It's a lot more capable at agentic tasks than I expected. I keep discovering new capabilities every day, it's crazy for 1B active parame
We've been adopting kernels in transformers to handle the device-specific optims, quants, or easyinstalls for popular kernels. We're opening this gradually with a focus on impactful kernels; if you're
I'm really upset about this: OpenAI's Codex Desktop had a 'Copy as Markdown' option for exporting full chat transcripts, but the feature vanished in an update a couple of days ago Genuinely my single
We need more coding and agent traces public sharing to build datasets and better open source models! Lots of people contributing already, you should share yours too! https://huggingface.co/datasets?se
AI safety can't happen behind closed doors! Super cool to see that the @AISecurityInst is releasing its evals, datasets, and models in the open on @huggingface, so researchers everywhere can scrutiniz
Agree! was talking about this with @havoyan just a few days ago. That's also the reason why so much of the value has been accruing to the frontier models in my opinion (cc @GavinSBaker) because if you
Artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships, and do not know from within what love, work, friendship or res
Fine-tune your first AI model today. Run GPT4o level model and run on your phone or laptop. @OpenBMB released 15M samples SFT dataset that you can use right now. (319GB high-quality post training data
NVIDIA has released a corrected version of DeepSeek-V4-Pro-NVFP4, a quantized model variant optimized for NVIDIA hardware using NV-FP4 (NVIDIA's 4-bit floating-point format). The model is available on
llama.cpp now has an official website: https://llama.app Our goal is to make local AI accessible to everyone, and improving the user experience is a big part of that. On the new landing page you’ll fi
love it! @ggerganov 's llama.cpp delivering! pibot is now running fully local, using parakeet for STT, qwen3-tts for TTS, and Qwen 3.6 as the local multi-modal LLM via llama.cpp. The STT and TTS infer
Most people know Hugging Face from its public models and datasets but few realize that 50% of the models and datasets stored on HF are private. This number has been increasing with buckets (our S3 alt
Most people training agentic LLMs with RL right now have a silently broken training loop and have no idea. Here's the trap: single-turn RL works beautifully. Clean curves, sane rewards, everything con
pibot is now running fully local, using parakeet for STT, qwen3-tts for TTS, and Qwen 3.6 as the local multi-modal LLM via llama.cpp. The STT and TTS inference engines are Rust/mlx-c based. Ported fro
Thanks for the call out and glad to see that our blog inspired the broader community! We at @FireworksAI_HQ have been using delta updates to scale RL globally for about a year. Including Composer 2 &
This week, I got our GitHub Actions to use @HuggingFace Jobs instead of the default GitHub CI runners, making workflows run on much more reliable CPUs or even on serverless GPU (that cost less than a
Noah McLaughlin created a 7-billion parameter language model and published it on Hugging Face, drawing a humorous parallel to how AI practitioners evaluate models using model cards by suggesting human
This post announces an 8-9 billion parameter language model optimized for efficiency and speed, trained on a very large token dataset, and positioned as accessible for users with limited GPU resources
just noticed today - the dataset is already past 1k+ downloads. opensource / openresearch ftw ! @evo__hq would be opensourcing as many datasets, evals and autoresearch runs as we can in our pursuit of
Local Reachy Mini conversations wireless looks like magic! You can bring your friend around the house and get WOW effect from anyone! 🔥 Thanks @andimarafioti for the blog post on how to set this up: h
Narrative violation: Open model use in Factory has more than 3x’d in the last month relative to closed models. By both total consumption and event count. Will be interesting to see what open vs closed
📢 New @heyjasper release ! 📢 MONET 🌸 : An Apache2.0 deduped and recaptioned dataset of 105M samples unlocking reproducible text-to-image research. Nano T2I 🖌️ : A codebase to train your own T2I model
NVIDIA has released GLM-5.1-NVFP4, a quantized version of the GLM-5.1 model, now available on Hugging Face. The model appears to use NVFP4 (NVIDIA's floating-point 4-bit) quantization format, designed
Spectacular work by @ClementDelangue and the team at @huggingface! The HF science team just made async RL weight sync ~100x cheaper on bandwidth, and you don't need a shared cluster anymore. The probl
The HF science team just made async RL weight sync ~100x cheaper on bandwidth, and you don't need a shared cluster anymore. The problem: every RL step, the trainer typically has to sync fresh weights
today is (potentially) a great day for the GPU poors if DiffusionBlocks works on fine-tuning existing models, then literally any reasonable consumer GPU can do LLM fine-tuning will make a video on thi
Today, we're releasing LFM2.5-8B-A1B, a device-optimized model designed to power real-life applications on phones, laptops, PCs, robots, and fast & lightweight server-side use-cases. > 8B MoE, 1.5B ac
We are starting to be quite bullish about getting in the data infrastructure business. I just cloned 68 TB (while I only have a 4TB local disk) to my @huggingface training bucket in 1 minute 55 second
We're releasing Paris 2.0, which, to our knowledge, is the world's first decentralized trained video generation model. We benchmarked it against a monolithic model trained on the same data and compute