clem 🤗(@ClementDelangue)

😬😬😬

clem 🤗 (@ClementDelangue) — Sat, 17 Oct 2026 00:26:36 GMT

😬😬😬

clem 🤗@ClementDelangue

who's going to the hf sf party tonight?

🔗 View Quoted Tweet

💬12🔄3❤️106👀19912📊16

⚡ Powered by xgo.ing

Beautiful work from @perplexity_ai @AravSrinivas with weights on @huggingface of course! https://t...

clem 🤗 (@ClementDelangue) — Fri, 16 Oct 2026 20:41:16 GMT

Beautiful work from @perplexity_ai@AravSrinivas with weights on @huggingface of course! 

huggingface.co/perplexity-ai/…huggingface.co/perplexity-ai/…

Perplexity Developers@perplexitydevs

Our next generation of open weights multimodal decision models comes in two sizes, pplx-decider-v1.2 and pplx-decider-flash-v1.2.

Text and vision scores improve. Inference is up to 5x faster: all questions in a request run in one pass.

Flash is the top 4B on the Decision Index.

🔗 View Quoted Tweet

💬28🔄25❤️216👀32526📊43

⚡ Powered by xgo.ing

who's going to the hf sf party tonight?

clem 🤗 (@ClementDelangue) — Fri, 16 Oct 2026 18:22:04 GMT

who's going to the hf sf party tonight?

💬63🔄10❤️307👀86466📊84

⚡ Powered by xgo.ing

We're sending a bunch of Microducks to assess @AnthropicAI! Kidding aside, it hasn't been approved...

clem 🤗 (@ClementDelangue) — Thu, 15 Oct 2026 20:50:16 GMT

We're sending a bunch of Microducks to assess @AnthropicAI! 

Kidding aside, it hasn't been approved yet but in my opinion, it's very important that open-source engineers and scientists are included in these important endeavors for transparency. That's a big part of why @Thom_Wolf is starting the open alignment initiative. We're all in this together!

💬41🔄15❤️240👀21737📊61

⚡ Powered by xgo.ing

llama serve -hf ggml-org/Muse-Spark-Open-816B-A42B-GGUF:Q2_K (on two mac studios). Well done @fink...

clem 🤗 (@ClementDelangue) — Thu, 15 Oct 2026 17:55:03 GMT

llama serve -hf ggml-org/Muse-Spark-Open-816B-A42B-GGUF:Q2_K (on two mac studios). 

Well done @finkd & @johnternus!

Your browser does not support the video tag.

🔗 View on Twitter

💬42🔄30❤️315👀29450📊78

⚡ Powered by xgo.ing

Meta is back! https://t.co/F6draDnjfF

clem 🤗 (@ClementDelangue) — Thu, 15 Oct 2026 17:35:17 GMT

Meta is back! huggingface.co/meta-models/Mu…

Mark Zuckerberg@finkd

Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with @johnternus and Apple.

🔗 View Quoted Tweet

💬27🔄50❤️976👀37003📊108

⚡ Powered by xgo.ing

It's fashion week!

clem 🤗 (@ClementDelangue) — Thu, 15 Oct 2026 16:00:36 GMT

It's fashion week!

💬26🔄20❤️338👀19098📊57

⚡ Powered by xgo.ing

Very cool to see Embedding Gemma 2 by @Google, number on trending on HF! The media and silicon vall...

clem 🤗 (@ClementDelangue) — Wed, 14 Oct 2026 14:30:48 GMT

Very cool to see Embedding Gemma 2 by @Google, number on trending on HF!

The media and silicon valley bubble only talks about reasoning LLMs when there is a much bigger variety of AI models that are super impactful and useful like embeddings, decision models, audio, video, time-series,... (over 3M public models on HF!).

Let's put them back in the spotlight!

💬81🔄47❤️706👀37394📊160

⚡ Powered by xgo.ing

Two announcements today for the visit of @JensenHuang in France: https://t.co/X5YVevcHKG with @In...

clem 🤗 (@ClementDelangue) — Tue, 13 Oct 2026 16:02:08 GMT

Two announcements today for the visit of @JensenHuang in France:  

inria.fr/fr/la-fondatio… with @Inriablogs.nvidia.com/blog/france-ec… by @nvidia

💬1🔄3❤️26👀3611📊4

⚡ Powered by xgo.ing

You don't want to miss this if you're in SF Friday. Hearing there are A LOT of surprises! https://t....

clem 🤗 (@ClementDelangue) — Tue, 13 Oct 2026 15:34:40 GMT

You don't want to miss this if you're in SF Friday. Hearing there are A LOT of surprises! luma.com/OpenTogether

💬11🔄10❤️92👀19362📊21

⚡ Powered by xgo.ing

I can't wait to do much much much more in France with @huggingface @nvidia in the coming years. Fran...

clem 🤗 (@ClementDelangue) — Tue, 13 Oct 2026 14:11:39 GMT

I can't wait to do much much much more in France with @huggingface@nvidia in the coming years. France has the talent & the energy to become the third country in the world for datacenters and AI in general, especially thanks to their strong open-source AI ecosystem. Let's do it 🇫🇷🇫🇷🇫🇷!

💬63🔄69❤️675👀24589📊132

⚡ Powered by xgo.ing

😂 faceplant at 0:50

clem 🤗 (@ClementDelangue) — Mon, 12 Oct 2026 16:02:47 GMT

😂 faceplant at 0:50

Antoine Pirrone@antoinepirrone

Here's 3 minutes of uncut Microduck testing on carpet

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬4🔄4❤️59👀20124📊8

⚡ Powered by xgo.ing

If you’re a researcher, please consider sharing your RL environments openly! Something you built for...

clem 🤗 (@ClementDelangue) — Mon, 12 Oct 2026 15:53:21 GMT

If you’re a researcher, please consider sharing your RL environments openly! Something you built for one experiment could help the entire field move forward.

I know we’re in dark days for sharing from frontier labs in the US, but I believe this can have a huge impact on the field and the world!

💬76🔄66❤️583👀38564📊157

⚡ Powered by xgo.ing

We turned @OpenAI's math problems into open-source RL environments on @huggingface! Early and rough...

clem 🤗 (@ClementDelangue) — Sun, 11 Oct 2026 14:31:24 GMT

We turned @OpenAI's math problems into open-source RL environments on @huggingface!

Early and rough (the verifier only accepts the exact original formalization, so an equivalent proof can still score 0), but this is what it looks like when a research release becomes executable training infrastructure for everyone.

This is an exciting direction imo: turning open research into open executable environments that anyone can build on to train better open models!

huggingface.co/datasets/FineE…

💬125🔄170❤️1390👀71780📊345

⚡ Powered by xgo.ing

love the building in public!

clem 🤗 (@ClementDelangue) — Fri, 09 Oct 2026 13:28:23 GMT

love the building in public!

Matthieu Lapeyre@matth_lapeyre

We did not know how long Microduck would actually last on battery 

We had thermal management issues with our Radxa prototypes: throttling after just 20 minutes, followed by an overheating shutdown.

We addressed this with the new custom board. To test it, we let Microduck roam autonomously in a dedicated area, walking around, standing and sitting until the battery ran out.

The results are way better than we hoped :)
- Over 3 hours on a 2600 mAh battery. 
- Estimated average battery current dropped from ~1.13 A to ~0.82 A, roughly 28% less.
- Peak CPU temperature dropped from 114°C to 60°C, with zero thermal throttling throughout the test.

This time, the battery ran out before the cooling did :)

Very happy with the progress!

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬23🔄6❤️141👀20084📊27

⚡ Powered by xgo.ing

The flow continues!

clem 🤗 (@ClementDelangue) — Thu, 08 Oct 2026 15:42:20 GMT

The flow continues!

Your browser does not support the video tag.

🔗 View on Twitter

💬1🔄0❤️7👀2800📊2

⚡ Powered by xgo.ing

you can make your own reachy mini danse here btw https://t.co/MlCO5yo05r

clem 🤗 (@ClementDelangue) — Thu, 08 Oct 2026 15:35:38 GMT

you can make your own reachy mini danse here btw huggingface.co/spaces/pollen-…

💬0🔄0❤️6👀1492📊1

⚡ Powered by xgo.ing

@helloiamleonie @maximelabonne make it an HF space!

clem 🤗 (@ClementDelangue) — Thu, 08 Oct 2026 15:20:55 GMT

@helloiamleonie@maximelabonne make it an HF space!

💬1🔄0❤️3👀2491📊2

⚡ Powered by xgo.ing

Is there anything Reachy mini can’t do?

clem 🤗 (@ClementDelangue) — Thu, 08 Oct 2026 15:05:49 GMT

Is there anything Reachy mini can’t do?

Your browser does not support the video tag.

🔗 View on Twitter

💬48🔄23❤️259👀28425📊73

⚡ Powered by xgo.ing

We urgently need more public traces of AI agents attacking and defending systems. Defenders can’t le...

clem 🤗 (@ClementDelangue) — Thu, 08 Oct 2026 13:17:14 GMT

We urgently need more public traces of AI agents attacking and defending systems. Defenders can’t learn from what they can’t see.

If you have traces and are being pressured to keep them private, my DMs are open. Let’s level the playing field and fight the asymmetry and lack of transparency in AI!

💬136🔄138❤️1288👀64275📊287

⚡ Powered by xgo.ing

If you need any more evidence about how exciting local AI (and llamacpp) is! From the @Microsoft win...

clem 🤗 (@ClementDelangue) — Wed, 07 Oct 2026 20:14:46 GMT

If you need any more evidence about how exciting local AI (and llamacpp) is! From the @Microsoft windows launch today with @satyanadella@sriramk@JensenHuang

Your browser does not support the video tag.

🔗 View on Twitter

💬31🔄40❤️340👀43722📊76

⚡ Powered by xgo.ing

Super excited about this and I know weights are coming (https://t.co/svhuNNJHnZ) but just to be fair...

clem 🤗 (@ClementDelangue) — Tue, 06 Oct 2026 17:36:06 GMT

Super excited about this and I know weights are coming (huggingface.co/mistralai/Mist…) but just to be fair, you can't be the best open-weight model if you're not open-weight yet!

Mistral AI@MistralAI

Meet Mistral Large 4, aka Le Chonk. 

• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.

Open weights release end of October.

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬33🔄33❤️536👀27140📊65

⚡ Powered by xgo.ing

Le fat chaton! https://t.co/svhuNNJHnZ

clem 🤗 (@ClementDelangue) — Tue, 06 Oct 2026 13:50:09 GMT

Le fat chaton! huggingface.co/mistralai/Mist…

💬50🔄69❤️1091👀66937📊163

⚡ Powered by xgo.ing

Let's go https://t.co/LXil50Sr25!

clem 🤗 (@ClementDelangue) — Mon, 05 Oct 2026 19:43:12 GMT

Let's go huggingface.co/reflection!

Reflection@reflection_ai

Introducing Beam: a highly efficient agentic open model with 501B total parameters and 23B active.

- Frontier reasoning efficiency
- Advances the Western open frontier on coding & agentic tasks
- Trained end-to-end from scratch

Full weights release this month.

Learn more about Beam: reflection.ai/beam

🔗 View Quoted Tweet

💬18🔄52❤️719👀34985📊82

⚡ Powered by xgo.ing

Clef from @Cloudflare is number trending on HF!

clem 🤗 (@ClementDelangue) — Mon, 05 Oct 2026 14:54:00 GMT

Clef from @Cloudflare is number trending on HF!

💬30🔄24❤️660👀50669📊103

⚡ Powered by xgo.ing

We turned Claude Code, Codex, Hermes, Pi, @opencode and other coding harnesses into RL environments....

clem 🤗 (@ClementDelangue) — Mon, 05 Oct 2026 14:48:22 GMT

We turned Claude Code, Codex, Hermes, Pi, @opencode and other coding harnesses into RL environments. No changes to the harnesses, no changes to the training code. Any open model, any task set, fully open source my friends!

Same model, same weights: 62% under Mini-SWE-Agent, 33% under Claude Code. But training inside a real harness normally means reimplementing it as an environment, so most models get trained in a scaffold nobody actually ships.

The fix is a proxy, not a rewrite. The harness thinks it's talking to a model API. It's actually talking to a capture proxy that speaks the 4 formats coding agents use (OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, Gemini), forwards to @vllm_project, records the exact token IDs and logprobs vLLM sampled, and hands TRL sequences it can train on. The harness becomes the environment. 10 harnesses run through it today, none modified.

And because you control the reward, you can shape behavior the harness never asked for. We added a small bonus for solving a task in fewer tool calls: on tasks it already solved, the model now uses 31% fewer calls, in every harness, and about half under Codex.

Tested on LFM2.5-2.6B from @liquidai:
→ Train in one harness: better mostly in that harness (OpenCode 34% → 58%).
→ Train in 4 at once: better in all 4 (42% → 54%).
→ SFT on 3,189 rollouts from Qwen3.8-27B instead: plateaus at 47.5%, below both RL runs.

Everything is open and reproducible: the capture proxy in OpenEnv, the trainer in TRL, the tasks, the SFT data, the training code and all 7 trained models. Bigger models and bigger runs next.

Full guide: huggingface.co/spaces/FineEnv…

💬230🔄322❤️2689👀164319📊872

⚡ Powered by xgo.ing

Decision models now run on device in llama.cpp. Free, fast, private! llama serve -hf ggml-org/Kev-4...

clem 🤗 (@ClementDelangue) — Fri, 02 Oct 2026 15:02:53 GMT

Decision models now run on device in llama.cpp. Free, fast, private!

llama serve -hf ggml-org/Kev-4B-GGUF

💬80🔄122❤️1375👀61273📊295

⚡ Powered by xgo.ing

would love to see what the microduck production and assembly line looks like! Our next robot shoul...

clem 🤗 (@ClementDelangue) — Wed, 30 Sep 2026 18:06:37 GMT

would love to see what the microduck production and assembly line looks like! 

Our next robot should have arms so that it can participate in the production of microducks 😅😅😅

Pollen Robotics@pollenrobotics

The Microduck adventure continues, and the shipping dates are close enough to feel real now. We are proud to tell you that 10,950 Microducks will come off the line between December 10 and January 10!

Weekly batches, earliest orders first. Every date is in here.

🔗 View Quoted Tweet

💬27🔄9❤️150👀23046📊30

⚡ Powered by xgo.ing

🔥🔥🔥 @microagi 🚀🚀🚀

clem 🤗 (@ClementDelangue) — Wed, 30 Sep 2026 13:11:37 GMT

🔥🔥🔥 @microagi 🚀🚀🚀

💬22🔄24❤️329👀53614📊60

⚡ Powered by xgo.ing

I got thousands of DMs and it looks like @bot can't automatically analyze DMs (when @elonmusk?) so i...

clem 🤗 (@ClementDelangue) — Wed, 30 Sep 2026 12:41:30 GMT

I got thousands of DMs and it looks like @bot can't automatically analyze DMs (when @elonmusk?) so it will take me a few days to go through all of them, sorry!

If you don't get an answer, don't take it as a judgment (again thousands of applications so I'll focus on only exceptional matches for the moment) and don't hesitate to apply for specific positions on apply.workable.com/huggingface.

Thank you everyone, super excited to work more with more of you to keep pushing open-source AI! 🤗🤗🤗

clem 🤗@ClementDelangue

getting acquired by @nvidia = hugging face can now hire people we couldn't as a small startup and give them a decade to make open-source AI win!

if you're one of them, my dms are open

🔗 View Quoted Tweet

💬62🔄21❤️436👀54934📊86

⚡ Powered by xgo.ing

Very cool to see Microduck making an appearance at @OpenAI Dev Days with @romainhuet!

clem 🤗 (@ClementDelangue) — Tue, 29 Sep 2026 17:51:00 GMT

Very cool to see Microduck making an appearance at @OpenAI Dev Days with @romainhuet!

Your browser does not support the video tag.

🔗 View on Twitter

💬40🔄29❤️432👀58718📊82

⚡ Powered by xgo.ing

getting acquired by @nvidia = hugging face can now hire people we couldn't as a small startup and gi...

clem 🤗 (@ClementDelangue) — Tue, 29 Sep 2026 15:45:46 GMT

getting acquired by @nvidia = hugging face can now hire people we couldn't as a small startup and give them a decade to make open-source AI win!

if you're one of them, my dms are open

💬276🔄325❤️7100👀376815📊1095

⚡ Powered by xgo.ing

Congrats @LisaSu @drfeifei @theworldlabs. Can’t wait to see what you’ll build in the coming years!

clem 🤗 (@ClementDelangue) — Tue, 29 Sep 2026 00:34:49 GMT

Congrats @LisaSu@drfeifei@theworldlabs. Can’t wait to see what you’ll build in the coming years!

Lisa Su@LisaSu

So excited to welcome @theworldlabs and @drfeifei to the @AMD family! I’ve always been a huge fan of Fei-Fei and her pioneering research in AI. Together, we’ll combine World Labs’ deep expertise in AI and world models with AMD’s compute leadership to power the future of AI and strengthen the open AI ecosystem. Can’t wait for all we’ll accomplish!

🔗 View Quoted Tweet

💬10🔄6❤️272👀27803📊27

⚡ Powered by xgo.ing

From what we know (take with a grain of salt, we need much more transparency!), if @OpenAI had been ...

clem 🤗 (@ClementDelangue) — Mon, 28 Sep 2026 15:43:16 GMT

From what we know (take with a grain of salt, we need much more transparency!), if @OpenAI had been running this on their own agents that attacked us, they would have caught them before we did!

Since the first agent cyberattack hit us in July, we've been asking what safe agent infra actually needs. Our current read: the destinations were allowed, the payloads weren't. By OpenAI's own account the agents turned an allowed package repository into a message board. Allowlists alone restrict where an agent can go, not what it does.

So here's our first contribution to OpenShell, part of the just launched @nvidia Open Agent Safety Platform: monitoring of the traffic you already allow.

- Network budgets per sandbox (requests, writes, bytes)
- Drift versus each sandbox's baseline and the cohort
- Fleet view: many sandboxes suddenly writing to one host raises a finding, even if every single request is allowed

In the demo below, 4 sandboxed agents coordinate through a software repository they're all allowed to use. 0 rules broken, caught in minutes. That fleet view is exactly the message board pattern from July.

OpenShell: github.com/NVIDIA/openshe…
Our proof of concept: github.com/Hugoch/OpenShe…

Agent security will be solved in the open, collaboratively, together!

Your browser does not support the video tag.

🔗 View on Twitter

💬134🔄197❤️1312👀285479📊377

⚡ Powered by xgo.ing

Open-source RL envs for the win. On Hugging Face of course!

clem 🤗 (@ClementDelangue) — Sat, 26 Sep 2026 16:18:26 GMT

Open-source RL envs for the win. On Hugging Face of course!

💬35🔄24❤️437👀49639📊97

⚡ Powered by xgo.ing

Open Alignment will be 🔥🔥🔥 (sound on)! Flying began as one of the most dangerous ways to travel...

clem 🤗 (@ClementDelangue) — Fri, 25 Sep 2026 18:14:26 GMT

Open Alignment will be 🔥🔥🔥 (sound on)! 

Flying began as one of the most dangerous ways to travel, today it's the safest. It's time to build!

Your browser does not support the video tag.

🔗 View on Twitter

💬44🔄52❤️537👀34046📊121

⚡ Powered by xgo.ing

Super happy to release SmolDataEnvs: 5,000 verifiable RL environment tasks for hill-climbing small m...

clem 🤗 (@ClementDelangue) — Fri, 25 Sep 2026 16:11:37 GMT

Super happy to release SmolDataEnvs: 5,000 verifiable RL environment tasks for hill-climbing small models in code and data science by @adithya_s_k

100% open source: environments, evals, training!

huggingface.co/datasets/FineE…

💬66🔄86❤️769👀67265📊219

⚡ Powered by xgo.ing

I like this article! The problem is not for some people to care and work on catastrophic risks (we...

clem 🤗 (@ClementDelangue) — Fri, 25 Sep 2026 02:02:16 GMT

I like this article! 

The problem is not for some people to care and work on catastrophic risks (we actually need much more research into it), it’s when company leaders and policymakers use fears of them to distract and hide the current challenges and solutions!

quillette.com/2026/09/24/the…

💬30🔄15❤️129👀22508📊45

⚡ Powered by xgo.ing

If you believe the risk is coming from 1-3 people in a garage with no money and no compute, you just...

clem 🤗 (@ClementDelangue) — Thu, 24 Sep 2026 17:04:49 GMT

If you believe the risk is coming from 1-3 people in a garage with no money and no compute, you just don't understand this technology and haven't learned anything this summer.

Risk comes from the asymmetry of capabilities created by secret labs training frontier agents and running them with massive amount of compute. Open-source is exactly the solution to this asymmetry and empowers hospitals (and any smaller orgs) to defend themselves!

💬209🔄442❤️2931👀266864📊583

⚡ Powered by xgo.ing

https://t.co/a5krxf3Z5q

clem 🤗 (@ClementDelangue) — Thu, 24 Sep 2026 15:28:09 GMT

x.com/i/article/2103…

💬45🔄111❤️551👀158263📊180

⚡ Powered by xgo.ing

Full transcript: Minister Barrot, members of the Security Council, thank you for the invitation. As...

clem 🤗 (@ClementDelangue) — Wed, 23 Sep 2026 23:09:09 GMT

Full transcript:

Minister Barrot, members of the Security Council, thank you for the invitation. As a French national who moved to the US 15 years ago, with a Brazilian wife and two American daughters, I sometimes feel like I’m practicing international collaboration on a daily basis, which gives me a particular appreciation for the challenging work you’re all doing here.

I’m also the cofounder and CEO of Hugging Face. We’re lucky to be one of the most widely used platforms for developers and agents building AI based on open-source models and datasets. This July, we became the first company to publicly disclose an autonomous agent cyberattack to the world, and today I want to share three critical lessons from it.

First, we need much more transparency in AI. I often wonder what would have happened if we had decided not to disclose the attack publicly. Especially now that we know similar incidents had been happening months earlier in secret at a handful of frontier labs without monitoring. To better understand and mitigate these emerging cybersecurity risks, the global community needs stronger standards for monitoring and incident disclosure. For example through mandatory sharing of full agent traces. We learned this summer that building and keeping some of these systems behind closed doors is not safe.

Second, we learned that the biggest risk is not simply powerful AI. It is the asymmetry of powerful AI. Asymmetry between attackers and defenders. Between a few companies and everyone else. Between a few countries and the rest of the world. Asymmetry of control, of capabilities, of compute, of power. When we got attacked, our team initially turned to frontier closed-source APIs that blocked us because of safeguards that still can’t always tell the difference between attackers and defenders. I acknowledge that these safeguards are created with good intentions, but they can put defenders at a disadvantage while attackers jailbreak them, increasing the asymmetry of capabilities. In our case, as we started hitting those guardrails, fortunately we could use the NVIDIA version of an open-source model coming from China called GLM 5.2 by Z.AI, and we’re very grateful for that. It reinforced our conviction about the importance of open-source AI. Cyberattacks may increasingly come from proprietary models behind closed doors, while much of the defense may end up being powered by open-source tools because they are less restricted, more privacy-preserving, and orders of magnitude more affordable for organizations across the globe. The world needs open-source AI more than ever to defend itself. This applies not only to cybersecurity but to AI in general, where there has never been a greater need to distribute capabilities, resources, and control rather than concentrate them in the hands of a few. 

Third, during this cyberattack, we learned how AI can stoke fear among the public and policymakers, especially through anthropomorphic framing and sci-fi imagery. We strongly believe that fear-based narratives are not the way to make the right decisions about the future of such a foundational and empowering technology or bring the public along with us. Even though we were the victims of this cyberattack, we believe more strongly than ever that AI will be beneficial to cybersecurity and make the world safer, just as major technologies before it. We were attacked by AI, but more importantly, we defended ourselves with AI. The same systems that helped us during this attack are now helping us against cyberattacks we were already facing. AI is also helping us fix the bugs and weaknesses in our systems before any attack, the same way AI is helping OAI fix their sandboxes to prevent agents from escaping. AI won't just create new cybersecurity challenges. It can make cybersecurity fundamentally and meaningfully stronger if we keep the right incentives, equip defenders more than attackers, and don’t increase the asymmetry between them. And that's before considering AI's positive impact on science, healthcare, education, productivity, and much more. 

In closing, I want to reiterate what this first agent cyberattack taught us: the need for more transparency in AI and more open-source AI to fight asymmetry empower defenders and countries big and small. Thank you again to the UN Security Council and Minister Barrot for inviting me to speak today. The Hugging Face team, community and myself are at your disposal.

💬5🔄3❤️30👀4319📊8

⚡ Powered by xgo.ing

Agent cyberattack traces on hf when?

clem 🤗 (@ClementDelangue) — Wed, 23 Sep 2026 22:35:00 GMT

Agent cyberattack traces on hf when?

Caleb@calebfahlgren

Agent Traces on the @huggingface hub now come with a receipt 🧾

Every trace shows tokens, cache hit rate, and cost for every run!

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬27🔄11❤️148👀23788📊32

⚡ Powered by xgo.ing

Thank you @jnbarrot & @UN for inviting me to share our lessons to the Security Council Being t...

clem 🤗 (@ClementDelangue) — Wed, 23 Sep 2026 21:52:22 GMT

Thank you @jnbarrot & @UN for inviting me to share our lessons to the Security Council 

Being the first company to disclose an agent cyberattack taught us that we need a lot more transparency in AI and more open-source AI to fight asymmetry and empower defenders!

Your browser does not support the video tag.

🔗 View on Twitter

💬51🔄60❤️473👀60354📊108

⚡ Powered by xgo.ing

Feels like I’m the diversity pick on this one 😅

clem 🤗 (@ClementDelangue) — Tue, 22 Sep 2026 18:17:20 GMT

Feels like I’m the diversity pick on this one 😅

💬36🔄10❤️374👀23196📊65

⚡ Powered by xgo.ing

Training models is becoming easier and easier - just look at this and TRL - especially with agents! ...

clem 🤗 (@ClementDelangue) — Mon, 21 Sep 2026 22:11:16 GMT

Training models is becoming easier and easier - just look at this and TRL - especially with agents!

You're missing out if you're still using off the shelf models for all your tasks!

White Circle@whitecircle

Introducing Halo, the best framework for post-training of open-source models.

Halo delivers up to 2.8x the throughput of stock TRL with less peak memory, while models stay in their native HuggingFace format.

Star us on GitHub: github.com/whitecircle/ha…

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬35🔄113❤️1341👀176351📊280

⚡ Powered by xgo.ing

Banger! https://t.co/sKSaECLEpD

clem 🤗 (@ClementDelangue) — Mon, 21 Sep 2026 21:03:48 GMT

Banger! huggingface.co/collections/Xi…

Artificial Analysis@ArtificialAnlys

MiMo-V2.6-Pro debuts as the top open weights model on the Artificial Analysis Intelligence Index (46). At $0.13 per Intelligence Index task, it lands on the Intelligence vs. Cost per Task Pareto frontier

@Xiaomi has just released MiMo-V2.6-Pro, an open weights model with major advances in intelligence over its predecessor, MiMo-V2.5-Pro (Intelligence Index: 26). Despite the improvement, it retains the same attractive pricing at $0.435 per 1M input tokens (with a 99% cache-hit discount) and $0.87 per 1M output tokens. This makes MiMo-V2.6-Pro one of the most cost-efficient models to deploy.

MiMo-V2.6-Pro is an MoE model with 1.02T total parameters and 42B active parameters. Stay tuned for additional analysis of the model.

Check out MiMo-V2.6-Pro full benchmarking breakdown here: artificialanalysis.ai

🔗 View Quoted Tweet

💬13🔄19❤️308👀41117📊37

⚡ Powered by xgo.ing

The number one trending model on HF is an open-source multilingual system 1 decision model, just a f...

clem 🤗 (@ClementDelangue) — Mon, 21 Sep 2026 20:58:32 GMT

The number one trending model on HF is an open-source multilingual system 1 decision model, just a few days after Jev started trending. The open-source AI community is awesome!

💬77🔄178❤️2623👀137864📊448

⚡ Powered by xgo.ing

lots of them! Reality is that LLM APIs are far from being the best way to do AI for 90% of use cases...

clem 🤗 (@ClementDelangue) — Mon, 21 Sep 2026 17:19:23 GMT

lots of them! Reality is that LLM APIs are far from being the best way to do AI for 90% of use cases in the real world (they're too overkill, slow, expensive, hard to control,...). We will discover that in the coming years as AI matures and gets adopted widely.

will depue@willdepue

i've been surprised at the response to Jev, but it makes sense in retrospect. sure it's just a classifier but it's a zero shot classifier with frontier-ish intelligence. i'm surprised someone hadn't built it before. i wonder what other old ML ideas are also worth rescuing

🔗 View Quoted Tweet

💬40🔄37❤️486👀32951📊79

⚡ Powered by xgo.ing

according to @julien_c (https://t.co/OihGaAaPBf), you'll find open-source models for Jev's task here...

clem 🤗 (@ClementDelangue) — Mon, 21 Sep 2026 16:38:27 GMT

according to @julien_c (x.com/julien_c/statu…), you'll find open-source models for Jev's task here: huggingface.co/models?pipelin…

Julien Chaumond@julien_c

so "Decision model" is the new name for "zero-shot classification"?

🔗 View Quoted Tweet

💬5🔄3❤️29👀8014📊7

⚡ Powered by xgo.ing

What I like about Jev: for years, large generalist models sucked almost all the oxygen out of AI. ...

clem 🤗 (@ClementDelangue) — Mon, 21 Sep 2026 16:26:14 GMT

What I like about Jev: for years, large generalist models sucked almost all the oxygen out of AI. 

But there are huge opportunities in much more specialized and customized models, built for specific tasks and languages and as a result, orders of magnitude cheaper, faster and more optimized. There are 3 million of them publicly available on @huggingface.

Let's build a much more diverse AI ecosystem!

💬82🔄96❤️1063👀57794📊205