Hugging Face(@huggingface)

LIVE STREAM: New Open Weight Model https://t.co/iBGsMXYfvr

Hugging Face (@huggingface) — Fri, 16 Oct 2026 14:00:32 GMT

LIVE STREAM: New Open Weight Model x.com/i/broadcasts/1…

💬39🔄18❤️154👀34460📊56

LIVE STREAM in minutes! Hear from the Reflection team about how they built Beam. Stream: https://t....

Hugging Face (@huggingface) — Fri, 16 Oct 2026 13:15:47 GMT

LIVE STREAM in minutes! Hear from the Reflection team about how they built Beam.

Stream: youtube.com/watch?v=dGhUFq…

Reflection@reflection_ai

Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners.

We're releasing:
- Weights, tokenizer, and reference implementation
- BF16, FP8 and NVFP4 weights and a DFlash2 drafter
- The open source harness we built and trained Beam on: Mirror CLI

Weights and model card on Hugging Face: huggingface.co/reflection

💬29🔄12❤️109👀35415📊32

Learn about how to get multipliers in local LLM speed-ups using speculative decoding in this livestr...

Hugging Face (@huggingface) — Wed, 14 Oct 2026 12:41:35 GMT

Learn about how to get multipliers in local LLM speed-ups using speculative decoding in this livestream with @mervenoyann, @MaartenGr and @_EldarKurticyoutube.com/watch?v=H4azmW…

💬83🔄38❤️200👀54833📊115

Help keep science open, try this out on Paper Pages: https://t.co/1cApLwjR7A

Hugging Face (@huggingface) — Thu, 08 Oct 2026 17:45:55 GMT

Help keep science open, try this out on Paper Pages: hf.co/papers

🍉 Abubakar Abid@abidlabs

🚨 The incentives to share scientific knowledge openly are crumbling.

Historically, researchers shared scientific knowledge for two primary reasons: (1) the prestige conferred by discovery, and (2) a sense of expanding the pool of human knowledge. Even within industry, scientific discovery was far too vast an endeavor for a single scientist or team to tackle alone. So scientists explored different frontiers, usually published their findings, and collectively moved the field forward.

However, as more scientific knowledge is uncovered simply by throwing massive compute at autonomous agents with minimal human intervention, these  incentives are breaking down: (1) discoveries are becoming commoditized. If anyone could have found the same result simply by running the same agentic prompt with enough compute, the prestige of discovery and the pride in one's work disappears, a shift we are already witnessing in software engineering. (2) the explosion of  low-quality, low-barrier publications makes genuine breakthroughs increasingly difficult to filter, discover, and appropriately credit. (3) and for companies, keeping discoveries proprietary is increasingly advantageous. Instead of publishing and waiting for the broader scientific community to build upon the work, an organization can simply turn its own compute back onto the problem to iterate faster internally.

In the short term, the incentives reward secrecy. But my goal -- and at Hugging Face, our goal -- is the sustainable, long-term expansion of human knowledge that survives individual, closed companies. So we have to build new tools that reward and facilitate open sharing of knowledge. If agents are accelerating discovery, we are going to figure out how to use agents and other tools to preserve scientific integrity. 

Starting with this: Paper Reproductions with ML Intern -- we’ve made it possible to take a published paper and task agents with independently reproducing its experiments, helping create more signals (e.g. open artifacts) that can help identify high-quality work.

Your browser does not support the video tag.

💬26🔄13❤️87👀33206📊29

The same model, with the same weights, scores 62% in one agent harness and 33% in another. @adithya...

Hugging Face (@huggingface) — Fri, 02 Oct 2026 14:51:01 GMT

The same model, with the same weights, scores 62% in one agent harness and 33% in another.

@adithya_s_k and the @huggingface team just released the ultimate guide to multi-harness RL, and it's one of the most practical RL write-ups this year, and everything open!

The trick is simple. Don't touch the harness. Point it at a proxy instead of the model. The proxy speaks all four API formats coding agents use (OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, Gemini). It records the exact token ids and logprobs vLLM sampled, and you train on that. You don't change a single line of Claude Code, Codex or OpenCode.

Results:
🔹 Trained across 4 harnesses at once, LFM2.5-2.6B by @liquidai went from 42% to 54%
🔹 31% fewer tool calls, thanks to a small bonus for solving tasks in fewer steps
🔹 Training in OpenCode alone took OpenCode from 34% to 58%, but the multi-harness model improved everywhere

They also tried the shortcut everyone reaches for: fine-tune on 3,189 successful rollouts from Qwen3.8-27B. Imitation plateaued at 47.5%, below both RL runs. Copying a bigger model doesn't get you there. Practice does.

The best part is that everything is open: the capture proxy in OpenEnv, the trainer in TRL, the tasks, the SFT data, the training code and all seven trained models.
Agents will run in dozens of harnesses. 

Now open models can be trained for each of them, by anyone.

Read it herehuggingface.co/spaces/FineEnv…9SLS

💬193🔄159❤️1259👀135674📊489

MCPs are making quite the comeback! You can now use them to bring your data to https://t.co/tLDUmb9U...

Hugging Face (@huggingface) — Thu, 01 Oct 2026 15:37:07 GMT

MCPs are making quite the comeback! You can now use them to bring your data to hf.co/chat?mode=ml-i…

💬60🔄26❤️187👀37824📊87

What would you like to see in Gradio? Drop your feature requests below 👇

Hugging Face (@huggingface) — Mon, 28 Sep 2026 19:07:19 GMT

What would you like to see in Gradio? Drop your feature requests below 👇

🍉 Abubakar Abid@abidlabs

5 years ago, we released @Gradio to make it easier to build frontends for AI apps. At that time, that was the main blocker for making AI more accessible to the community. But this is no longer a challenge. So we asked ourselves and other developers: "what are the main challenges that you face now when building and using AI apps" – and we are now working hard to redesign Gradio from scratch to solve these issues!

💬54🔄10❤️73👀42982📊44

Training Agents 4: From reward functions to environments. https://t.co/9VSrFcBiPU

Hugging Face (@huggingface) — Thu, 10 Sep 2026 15:00:22 GMT

Training Agents 4: From reward functions to environments. x.com/i/broadcasts/1…

💬153🔄31❤️190👀102307📊160

🤗💚 https://t.co/coHXLgzNXh

Hugging Face (@huggingface) — Thu, 03 Sep 2026 13:47:02 GMT

🤗�blogs.nvidia.com/blog/nvidia-to…Xh

Your browser does not support the video tag.

💬261🔄597❤️5934👀499181📊944

We use Inkling-Small to turn paper abstracts into quick, useful summaries. open weights × open scie...

Hugging Face (@huggingface) — Wed, 26 Aug 2026 21:34:18 GMT

We use Inkling-Small to turn paper abstracts into quick, useful summaries.

open weights × open science 🤝

💬86🔄63❤️411👀169474📊132

We've just surpassed 3 million models on the Hub 🤗 the community is accelerating towards an open, ...

Hugging Face (@huggingface) — Tue, 18 Aug 2026 11:17:26 GMT

We've just surpassed 3 million models on the Hub 🤗

the community is accelerating towards an open, distributed future where open AI is everywhere, for everyone 🚀

Your browser does not support the video tag.

💬135🔄154❤️1197👀148552📊272

The State of Open Models, Summer 2026 ☀️ frontier models are getting larger, but small models still ...

Hugging Face (@huggingface) — Fri, 14 Aug 2026 16:28:41 GMT

The State of Open Models, Summer 2026 ☀️ frontier models are getting larger, but small models still dominate real-world usage. Qwen leads local inference, followed by Gemma. AI agents are becoming a major force on the Hub

Full picture on the blog 🤗huggingface.co/blog/state-of-…H

💬48🔄63❤️261👀85580📊89

would be even more popular than @UseCorgi cafe

Hugging Face (@huggingface) — Wed, 12 Aug 2026 12:48:33 GMT

would be even more popular than @UseCorgi cafe

Jess Fields@jessalanfields

Qwen 3.8 27b is being released at 11:00AM EST on Wednesday

Folks are lined up for miles to see what this fancy new model is capable of

💬27🔄10❤️118👀36808📊29

How AI agents reproduced ICML 2026 papers https://t.co/gReZNlWh7m

Hugging Face (@huggingface) — Fri, 07 Aug 2026 16:00:48 GMT

How AI agents reproduced ICML 2026  papers x.com/i/broadcasts/1…

💬20🔄11❤️69👀32052📊32

We're working with the Open Secure Alliance on guidelines for incident learning, to collectively dev...

Hugging Face (@huggingface) — Tue, 04 Aug 2026 21:53:10 GMT

We're working with the Open Secure Alliance on guidelines for incident learning, to collectively develop better review, disclosure, and controls of and for security incidents.

NVIDIA@nvidia

Security is strongest when organizations collaborate openly and move quickly. 

Now over 120 members strong, the Open Secure AI Alliance is sharing new open source security contributions, including new proposed SAFE guidelines, for turning confidential incident findings into stronger protection across the ecosystem.

💬18🔄9❤️56👀29992📊18

Pro tip: add in a "Sign-In with Hugging Face" OAuth button to your website/app so that community mem...

Hugging Face (@huggingface) — Wed, 29 Jul 2026 16:23:22 GMT

Pro tip: add in a "Sign-In with Hugging Face" OAuth button to your website/app so that community members can easily share their email, create model/dataset repos, store data in Buckets, or even kick off GPU-backed Jobs!

huggingface.co/docs/hub/en/oa…

💬24🔄16❤️110👀33923📊37

Training Agents 3: Learn how to do train a local/ open weight agent with reinforcement learning htt...

Hugging Face (@huggingface) — Tue, 28 Jul 2026 15:15:31 GMT

Training Agents 3: Learn how to do train a local/ open weight agent with  reinforcement learning x.com/i/broadcasts/1…

💬17🔄42❤️259👀32085📊82

AI security improves when organizations share research, tools and real-world experience. We’re join...

Hugging Face (@huggingface) — Mon, 27 Jul 2026 12:29:48 GMT

AI security improves when organizations share research, tools and real-world experience.

We’re joining industry leaders, including @NVIDIA, in the Open Secure AI Alliance to help organizations identify and address software vulnerabilities and strengthen critical systems.

Learn more: nvda.ws/4pD8Fc5

💬94🔄211❤️1823👀95613📊315

free the parameters 🤗

Hugging Face (@huggingface) — Sun, 26 Jul 2026 10:59:35 GMT

free the parameters 🤗

MiniMax (official)@MiniMax_AI

Open weights.
Open research.
Open innovation.🫶

Marching for an open future.🤍

Your browser does not support the video tag.

💬19🔄20❤️453👀69591📊50

Local AI 201 https://t.co/AgKvdJMqWv

Hugging Face (@huggingface) — Tue, 21 Jul 2026 16:02:30 GMT

Local AI 201 x.com/i/broadcasts/1…

💬26🔄16❤️117👀128207📊46

https://t.co/D5tKPltiWP

Hugging Face (@huggingface) — Sun, 19 Jul 2026 16:30:10 GMT

youtube.com/watch?v=90k2HR…

💬2🔄12❤️83👀22105📊27

Join us this Tuesday to learn more on Local AI, from software to hardware 🤗 We'll be joined by @Th...

Hugging Face (@huggingface) — Sun, 19 Jul 2026 16:23:30 GMT

Join us this Tuesday to learn more on Local AI, from software to hardware 🤗

We'll be joined by @TheAhmadOsman & @MikeBradleyAI covering hardware setups & local inference with live demo, and @alexocheema & @0xSero on picking model for your hardware, model compression and REAPs 🔥

Set your reminders to not miss out! 🔔

💬37🔄94❤️482👀204093📊159

New open weight model drop! https://t.co/TVdbLBZ8W0

Hugging Face (@huggingface) — Wed, 15 Jul 2026 18:28:16 GMT

New open weight model drop! x.com/i/broadcasts/1…

💬10🔄11❤️137👀60798📊27

Build Small Winners Reveal https://t.co/sD55EpWrnu

Hugging Face (@huggingface) — Fri, 10 Jul 2026 09:04:40 GMT

Build Small Winners Reveal x.com/i/broadcasts/1…

💬6🔄4❤️65👀39935📊14

Training Agents 2: Live tutorial on model distillation for training custom agents. https://t.co/tRqw...

Hugging Face (@huggingface) — Tue, 07 Jul 2026 15:00:34 GMT

Training Agents 2: Live tutorial on model distillation for training custom agents. x.com/i/broadcasts/1…

💬21🔄32❤️241👀37078📊88

Welcome to Open Source AI: Run Your Own Models Locally https://t.co/XtkDdgeBOP

Hugging Face (@huggingface) — Thu, 25 Jun 2026 15:00:25 GMT

Welcome to Open Source AI: Run Your Own Models Locally x.com/i/broadcasts/1…

💬18🔄55❤️346👀50997📊128

Build Small with Modal https://t.co/2rs4dOi5uE

Hugging Face (@huggingface) — Thu, 11 Jun 2026 17:29:29 GMT

Build Small with Modal x.com/i/broadcasts/1…

💬6🔄13❤️37👀13737📊15

Build Small Hackathon: OpenBMB https://t.co/yy1llS53uK

Hugging Face (@huggingface) — Wed, 10 Jun 2026 16:01:38 GMT

Build Small Hackathon: OpenBMB x.com/i/broadcasts/1…

💬6🔄9❤️34👀9825📊12

Build Small Hackathon https://t.co/SPUEZ049RM

Hugging Face (@huggingface) — Fri, 05 Jun 2026 16:01:34 GMT

Build Small Hackathon x.com/i/broadcasts/1…

💬10🔄8❤️62👀12821📊19

Nemotron 3 Ultra!

Hugging Face (@huggingface) — Thu, 04 Jun 2026 14:38:54 GMT

Nemotron 3 Ultra!

💬16🔄21❤️500👀35369📊63

Ideogram just released their latest and best v4 image model open weights State of the art and open ...

Hugging Face (@huggingface) — Wed, 03 Jun 2026 16:13:38 GMT

Ideogram just released their latest and best v4 image model open weights

State of the art and open weights go well together 🤗

Model:huggingface.co/ideogram-ai/id…D
Demo:huggingface.co/spaces/multimo…y 
x.com/ideogram_ai/st…0

Ideogram@ideogram_ai

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 and the API today.

Your browser does not support the video tag.

💬27🔄84❤️784👀89958📊146

Ideogram just released their latest and best v4 image model open weights State of the art and open ...

Hugging Face (@huggingface) — Wed, 03 Jun 2026 16:03:57 GMT

Ideogram just released their latest and best v4 image model open weights

State of the art and open weights go well together 🤗
x.com/ideogram_ai/st…0

Ideogram@ideogram_ai

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 and the API today.

Your browser does not support the video tag.

💬0🔄0❤️0

Hugging Face (@huggingface) — Mon, 01 Jun 2026 17:11:12 GMT

clem 🤗@ClementDelangue

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
- Current #1 trending model on HF with LocateAnything and #5 trending with PiD
- Announced that they're adopting the @linuxfoundation OpenMDW framework
- Released Cosmos 3, Omnimodal World Models for Physical AI & Alphamayo 2 Super, an open model for autonomous driving
- Announced the release soon of Nemotron 3 & work on Nemotron 4

Thank you @nvidia for all the work you're doing for the ecosystem and open-source AI. Can't wait for the next few months!

💬15🔄13❤️250👀23957📊32

We've just hit 1M open datasets on the Hugging Face Hub 🎉 Open models need open data. Today we hit...

Hugging Face (@huggingface) — Tue, 12 May 2026 15:26:10 GMT

We've just hit 1M open datasets on the Hugging Face Hub 🎉

Open models need open data. Today we hit that milestone, together with the most incredible community in AI! 🤗

Onwards to the next million 🚀

Your browser does not support the video tag.

💬31🔄75❤️617👀59126📊109

🐳 is back!

Hugging Face (@huggingface) — Fri, 24 Apr 2026 07:06:35 GMT

🐳 is back!

DeepSeek@deepseek_ai

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now chat.deepseek.com1Dl via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Repohuggingface.co/deepseek-ai/De…kYtp
🤗 Open Weighuggingface.co/collections/de…i7SDM

1/n

💬24🔄95❤️1773👀68338📊157

new open coding sota 🤗

Hugging Face (@huggingface) — Mon, 20 Apr 2026 15:52:29 GMT

new open coding sota 🤗

Kimi.ai@Kimi_Moonshot

Meet Kimi K2.6: Advancing Open-Source Coding

🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2)

What's new:
🔹Long-horizon coding  - 4,000+ tool calls, over 12 hours of continuous execution, with generalization across languages (Rust, Go, Python) and tasks (frontend, devops, perf optimization).
🔹Motion-rich frontend - Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D.
🔹Agent Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from K2.5's 100 / 1,500). One prompt, 100+ files. 
🔹Proactive Agents - K2.6 model powers OpenClaw, Hermes Agent, etc for 24/7 autonomous ops.  
🔹Claw Groups (research preview) - bring your own agents, command your friends', bots & humans in the loop.
-
K2.6 is now likimi.combwktG0 in chat mode and agent mode. 
For production-grade coding, pair K2.6 with Kimi kimi.com/codeJKyGCY
- 
�platform.moonshot.aikbOwCN4
🔗 Teckimi.com/blog/kimi-k2-6WvgIQSS3
🔗 Weightshuggingface.co/moonshotai/Kim…e0hjs2RTP

💬22🔄38❤️650👀50849📊72

llama-server -hf ggml-org/gemma-4-26b-a4b-it-GGUF:Q4_K_M openclaw onboard --non-interactive \ --a...

Hugging Face (@huggingface) — Sat, 04 Apr 2026 00:22:02 GMT

llama-server -hf ggml-org/gemma-4-26b-a4b-it-GGUF:Q4_K_M

openclaw onboard --non-interactive \
  --auth-choice custom-api-key \
  --custom-base-url "http://127.0.0.1:8080/v1" \
  --custom-model-id "ggml-org-gemma-4-26b-a4b-gguf" \
  --custom-api-key "llama.cpp" \
  --secret-input-mode plaintext \
  --custom-compatibility openai \
  --accept-risk

💬36🔄92❤️971👀98154📊275

Let's go!

Hugging Face (@huggingface) — Thu, 02 Apr 2026 16:25:07 GMT

Let's go!

💬8🔄44❤️538👀22537📊69

Model weights are here: https://t.co/rQlfP51Db7!

Hugging Face (@huggingface) — Fri, 27 Mar 2026 00:43:48 GMT

Model weights are here: huggingface.co/chromadb/conte…!

Chroma@trychroma

Introducing Chroma Context-1, a 20B parameter search agent.

> pushes the pareto frontier of agentic search

> order of magnitude faster

> order of magnitude cheaper

> Apache 2.0, open-source

Your browser does not support the video tag.

💬8🔄12❤️240👀41428📊37

do the right thing anon!

Hugging Face (@huggingface) — Tue, 24 Mar 2026 21:56:16 GMT

do the right thing anon!

SemiAnalysis@SemiAnalysis_

CLAUDE OPUS 5 ALERT: I was walking around San Francisco Hayes Valley today and saw an sidewalk this USB key saying it contains the opus 5 weights. Should i plug it into my computer & upload the weights to huggingface or should I give this USB key to @jekbradbury@aj_kourabi in case it is their USB?

💬5🔄6❤️96👀30772📊12

https://t.co/QLPgege4CI

Hugging Face (@huggingface) — Fri, 13 Mar 2026 12:50:42 GMT

huggingface2.notion.site/Hugging-Face-B…

💬0🔄1❤️41👀12511📊9

Seeing the worldwide demand we are kicking off global applications for Hugging Face Builders! If yo...

Hugging Face (@huggingface) — Fri, 13 Mar 2026 12:50:30 GMT

Seeing the worldwide demand we are kicking off global applications for Hugging Face Builders!

If you're passionate about open AI and love bringing people together, this is your invitation to lead ✉️

Learn more about the program and apply to become a Builder ➡️

💬12🔄28❤️243👀71397📊65

We are sponsoring Gemini hackathon with Cerebral Valley, see you this weekend!

Hugging Face (@huggingface) — Thu, 12 Mar 2026 15:46:33 GMT

We are sponsoring Gemini hackathon with Cerebral Valley, see you this weekend!

Cerebral Valley@cerebral_valley

Bonjour, Paris — it’s been a minute, but we’re back 🇫🇷

We're bringing Gemini 3 to town on March 14 with @GoogleDeepMind!

Join us to build and ship on Google’s most intelligent model yet, and compete for:

💰 $100,000 in Gemini API credits 
🎙️ An exclusive 30-minute virtual call with the founders of Google’s @AIFuturesFund

Register below👇

Your browser does not support the video tag.

💬1🔄19❤️188👀42363📊22

Learn more and apply from the link below🤗 https://t.co/QLPgege4CI

Hugging Face (@huggingface) — Thu, 12 Mar 2026 12:52:04 GMT

Learn more and apply from the link below🤗huggingface2.notion.site/Hugging-Face-B…I

💬0🔄1❤️6👀8956📊3

Hugging Face Builders is a global community program that puts local leaders at the center of the ope...

Hugging Face (@huggingface) — Thu, 12 Mar 2026 12:51:31 GMT

Hugging Face Builders is a global community program that puts local leaders at the center of the open-source AI movement 🤗

If you're passionate about open AI and love bringing people together, this is your invitation to lead ✉️
Apply for to build the Paris chapter today ➡️

💬10🔄20❤️123👀24628📊31

Read our blog to learn more 🤗 https://t.co/asj0iZulGe

Hugging Face (@huggingface) — Tue, 10 Mar 2026 17:53:41 GMT

Read our blog to learn more 🤗huggingface.co/blog/storage-b…e

💬0🔄4❤️22👀9464📊5

🪣 We just shipped Storage Buckets: S3-like mutable storage, cheaper & faster Git falls short for ...

Hugging Face (@huggingface) — Tue, 10 Mar 2026 17:53:08 GMT

🪣 We just shipped Storage Buckets: S3-like mutable storage, cheaper & faster 

Git falls short for everything on high-throughput side of AI (checkpoints, processed data, agent traces, logs etc) 

Buckets fixes that: fast writes, overwrites, directory sync 💨

All powered by Xet dedup so successive checkpoints skip the bytes that already exist ➡️

💬19🔄68❤️392👀69095📊92