
Pinned Tweet
13 months ago, I joined @bfl_ml as first the business hire among 14 researchers who pioneered Latent Diffusion, Stable Diffusion & FLUX.1.
Today, we announced our $300M Series B at a $3.25B valuation.
@robrombach, @andi_blatt, @pess_r, @AxSauer, @timudk & @sumith1896 founded BFL to make visual AI controllable, expressive, and powerful for anyone bringing ideas to life.
What makes this team different isn't just the technical firepower. It's the combination of deep research conviction and genuine care for building something exceptional that lasts. We are a team of 50 where everyone operates with high agency, warm intensity, and treats this like the most important work of their career.
Because for many of us, BFL is a destination company.
Within the last year, we defined what open frontier visual AI should look like and set a new standard for controllable image editing. We have gone from zero revenue to significant scale with Fortune 500 adoption, with FLUX being embedded in high‑value, high‑volume production workflows. We rival some of the largest companies in history in an enormous market and are just getting started.
Massive thank you to all our investors who continue backing our vision. I am incredibly proud of us as a team.
If you want high ownership, solve hard problems, and the chance to do the most important work of your career - join us into the new. 🌲
sarah retweeted
Black Forest Labs is on the inaugural 2026 AI 100.
Thanks @NewcomerMedia @BessemerVP for the recognition.
Join us. Links in 🧵
A great product doesn’t guarantee attention. A great launch doesn’t guarantee belief
(1) Most people think a viral launch is a good video and a lucky tweet. It’s not.
(2) It’s weeks of work: a story people actually care about, sharp copy, strong creative, a community that wants to share it, the right partners, and enough budget to give it a chance.
(3) AI made content cheap. Attention got expensive.
(4) The new GTM advantage isn’t making more content.
It’s having a story worth retelling.
2
17
658
Hiring brilliant, self directed people doesn’t make leadership unnecessary. Flatter orgs raise the bar.
Autonomous ICs still need clear priorities, useful feedback and someone invested in their growth. Leaders are still accountable for making that happen.
You can delegate decisions. You can’t delegate away responsibility for the environment people work in.
1
14
808
AI model launches are becoming Taylor Swift concerts: cryptic teasers, cinematic launch videos, fans decoding every frame. Leaderboards are the charts. API access is the presale. Somewhere, a vibe marketer is having their Eras Tour moment
6
625
skill gets you in the door, but giving a shit is what turns good work into generational work. you can’t teach giving a shit
On Friday I had a WOW moment watching my team work. One of the best feelings in my job is hiring someone who is simply better at it than I am.
It isn’t always comfortable. There’s a quiet voice that asks: if they’re better than you, what’s left for you? Get over it. In a fast growth company, listening to that voice makes you the bottleneck. The job is no longer just to be the best at the work. It’s to secure the superstars who are, and make sure they’re seen.
In a talent market this fierce, hiring exceptional people is only half the job. The other half is recognising your superstars before someone else does it for you. If the fastest route to promotion is an offer from somewhere else, we’re teaching our best people to go looking.
A counteroffer should never be the first serious conversation about someone’s value.
15
859
Hi from the forest. Two exciting roles have just opened up on our marketing team @bfl_ai
I’m incredibly proud of the exceptional work this team has done over the past year. Every person has raised the bar, and now we’re looking for people who will raise it again.
We’re small, moving fast and building at the frontier, with a lot of exciting work ahead. We care about quality, taste and sound judgement.
(1) Field Marketing Manager
Own our events, from booths and demos to exec dinners and side events, and build the co-marketing engine with our customers and partners. You’ll work closely with GTM, Partnerships and our creative technologists. If you think like a seller who can run special events, apply jobs.ashbyhq.com/black-fores…
(2) Brand Designer
Shape how Black Forest Labs shows up visually, online and in person, across launches, campaigns and events.
jobs.ashbyhq.com/black-fores…
If this sounds like you or someone you know, tag them below or drop me a message
sarah retweeted
The most built-upon model on the Hugging Face Hub isn’t Llama or Qwen.
It’s FLUX.1-dev from Black Forest Labs, a German lab: 43k derivative models.
Europe trained ~8% of the open-model compute on the Hub, but ~18% of what people build on top of it is European.
Build or surrender? Build, and build open. See you at GTC Berlin this week
sarah retweeted
The team Just got set up at GTC Berlin. You should come over to the @mimicrobotics x @bfl_ai booth, it's really good!
sarah retweeted
Introducing FLUX 3 Max.
Fastest in the west. Fast enough to try anything. And everything.
Generate a 10-second video in 12 seconds.
Tied for #1 in image-to-video. 22× faster than Seedance 2.5 at the same quality.
FLUX 3 is a model designed for customization. @fal adapted and optimized it to create FLUX 3 Max.
Calvin Klein dressed the models. We are letting you tailor this one
Meet Flux Image Klein: 5.6B parameters, open weights, runs on a single consumer GPU
FLUX 3 Image Klein is here.
Open Weights with the quality of a large frontier model. Customizable and fast.
5.6B parameters and runs on a single consumer GPU.
Native 4K, bounding boxes, and pixel-perfect editing.
6
20
764
GTC Berlin x BFL 🤖 🌲
join us for Afterglow on Wednesday 21st together with our friends @FactoryAI @modal @NVIDIAAI
See you there 👋
GTC Berlin, next week.
The team will be on the ground, come find us.
FLUX 3 Action at the @Nebius booth. FLUX 3 Image and Video at the NVIDIA booth, and a @mimicrobotics robot running FLUX-mimic.
@robrombach will be speaking on stage on Oct 22, 9AM for a panel and 4PM for a keynote.
We are hosting Afterglow, on Oct 21 with @FactoryAI, @modal and @NVIDIAAI, link in thread.
sarah retweeted
So far, europe2031.ai has got attention because the bad parts arrived early. Now a good part is arriving early.:
For January 2028, we wrote: "One of the [...] research bets – world models that can be used to train robots – is producing genuinely promising work that is getting international attention."
@bfl_ai did it in 2026, from a private lab in Freiburg.
The epilogue considered this "the real opening Europe had in robotics and industrial AI".
Watch this company very closely.
sarah retweeted
FLUX 3 is the strongest world model in the world.
FLUX 3 is #1 on Physics-IQ, @GoogleDeepMind's benchmark for physical world understanding in video models.
How Physics-IQ works: given the start of a real filmed experiment, can a model predict what physically happens next? Tests span fluid dynamics, optics, solid mechanics, magnetism and thermodynamics, covering how liquids flow, light behaves and objects interact.
In most cases, we substantially outperform all other frontier models, Seedance 2.5, MiniMax H3, Gemini Omni 1.1 Flash, Veo 3.1 and Sora 2, Cosmos3.
The FLUX 3 base model outperforms all other foundation models, while FLUX 3 paired with a physics verification layer scores even higher.
The FLUX 3 backbone, with its strong physics understanding, gives FLUX 3 Action it’s strong action prediction for robotics, while also giving FLUX 3 Video physically accurate motion when you’re creating content, making ads, films and product shots.
sarah retweeted
We made a new & nice model: FLUX 3 Image is out ✨
Built on the multimodal FLUX 3 backbone, it's extremely strong at multi-turn editing, generates natively at up to 4K, and supports bounding boxes for precise localized editing and generation.
An open-weight variant is coming.
Introducing FLUX 3 Image.
Control every pixel.
Make precise multi-turn edits without changing any other pixel.
Lay out the image exactly how you want using bounding boxes.
Generate in up to 4K to preserve details.
Use up to 10 references to compose an image.
Commercial Weights available for companies running image generation at scale.
Open Weights version of FLUX 3 Image is launching in the coming weeks.
sarah retweeted
Introducing FLUX 3 Image.
Control every pixel.
Make precise multi-turn edits without changing any other pixel.
Lay out the image exactly how you want using bounding boxes.
Generate in up to 4K to preserve details.
Use up to 10 references to compose an image.
Commercial Weights available for companies running image generation at scale.
Open Weights version of FLUX 3 Image is launching in the coming weeks.
🌲 ❤️ 🤖
Introducing FLUX 3 Action.
An open weights 7B World Action Model that achieves first place on the RoboLab benchmark.
It outperforms the previous best open model by 6.1 percentage points while using 56% fewer parameters and running up to 3.95x faster.
FLUX 3 Action removes the usual trade-off between world action model performance and VLA speed: it still predicts video and actions together, but plans more than twice as far ahead and runs faster per second of robot motion than the strongest open VLA.
Teams can fine-tune FLUX 3 Action on their own demonstrations to create policies for a particular robot and task. Together with @nvidia, we also integrated FLUX 3 Action natively into @huggingface's LeRobot, with fine-tuning recipes included and edge deployment on NVIDIA Jetson.
Beyond robotics, we’re also seeing promising results training task-specific policies for acting in simulated environments like gaming, controlling a vehicle, computer use, and wherever else a model needs to understand a visual environment and then choose what to do next.
FLUX 3 Action builds on the same image, video, and audio pretraining as FLUX 3, but uses a smaller architecture designed for practical deployment. In midtraining, we trained the model to predict actions and future frames together.
We’re releasing the weights, code, fine-tuning recipe, benchmarks, and reproducible examples so researchers and developers can build on the model with their own robots, environments, and tasks (see below).





















