sane

Dev and AI art

[amit @ CVPR @gravicle]

Uni-1.1 API is now live and creates a new pareto frontier - thinking images at the cost and efficiency of old school diffusion models. Uni-1.1 is a model designed for professional work, which continues to be difficult to capture in public benchmarks. Give it a try in Luma or your favorite genai tool!
incredibly proud of the Luma team for reaching the frontier and there's a lot more to come!

[William Shen @shenbokui]

🚀 UNI-1 debuts us as the best lab not named @OpenAI / @GeminiApp. Not bad for our first generation of unified image model!

Interestingly, with the current update, GPT Image 2’s ELO score is now 110 points lower than before. not sure what happened… cc @kenjihata @BoyuanChen0

Sharing a few genuinely GPT-Image-2-level generations from UNI-1 below 🤩

“Generate a news website page from the year 2036, featuring relevant news stories and ad blocks designed not for humans, but for AI agents who have evolved into distinct personalities. Both the website and all the advertisements featured on it should be in English.”

[sane @sane_codes]

turns out the embedding space has beautiful structure, check it out

[William Shen @shenbokui]

New viz mode drops.
Genuinely the coolest visualization I have seen.
We visualize UNI-1's latent embedding space as a 4D hypersphere using SO(4) rotation.
You can see how nearest neighbors move together (second half of the video).
Implemented by @sane_codes
Embedding is using siglip2 by @mtschannen @XiaohuaZhai @giffmana

[William Shen @shenbokui]

Most image models are good at one thing.
Uni-1 has been good at everything we've thrown at it.

Our team generated thousands of images leading up to Uni-1 launch. We embedded them all into a single map where visual similarity determines proximity. The result speaks for itself.

[William Shen @shenbokui]

UNI-1 is intelligent, directable, cultured. Incredible range it can do.

Incredibly proud of the world-class team building a world-class model.
It’s a daunting task to go up against industry giants like Deepmind/OpenAI/Bytedance.

More to come! API, technical report, model card…

Come join us!

[Luma @LumaLabsAI]

Uni-1 is here! A new kind of model that thinks and generates pixels simultaneously. Less artificial. More intelligent.

[amit @ CVPR @gravicle]

We launched Luma Agents last week.

ServicePlan — the largest independent agency in the world — has switched all their pitching to Luma Agents. A team of 3 at Mazda made a film spanning 40 years of the MX5. Adidas is generating campaigns before a stitch is sewn.

Here's what's going on:

[Jiaming Song @baaadas]

Excited to introduce Uni-1, our new *unified* multimodal model that does both understanding and generation: https://t.co/VkgMNnYtZv

TLDR: I think Uni-1 @LumaLabsAI is > GPT Image 1.5 in many cases, and toe-to-toe with Nano Banana Pro/2. (showcase below)

[Luma @LumaLabsAI]

Introducing Uni-1, Luma’s first unified understanding and generation model, our next step on the path towards unified general intelligence.
https://t.co/QjdrnYoWe5

[Luma @LumaLabsAI]

Introducing Luma Agents. Creative agents that make you prolific. You set the direction. They build with you, seeing what you see and helping teams explore further, iterate faster, and watch ideas multiply.

[Luma @LumaLabsAI]

“SOUL CODE” - Ep. 1
In the year 2099, humans can upload into machines, but it’s illegal. Where minds are stored, traded, and stolen, one street-level enforcer searches for what was taken from him. A recovered memory shard pulls him deep into a city filled with deadly permanence.

[Anastasia Fomina @exienator]

У меня есть свой небольшой музыкальный проект.
И вчера у меня вышел новый трек.
Приглашаю послушать.

[Luma @LumaLabsAI]

Choose your character. Reimagined with Ray3 Modify in Dream Machine.

[Luma @LumaLabsAI]

Stop guessing. Start directing. Ray3 Modify is now in Dream Machine. Edit and reimagine videos with all-new precise keyframe and character reference controls. Your vision, reimagined. Supercharge your production with rapid retouching, precise element swapping, and scene redesign.

[Luma @LumaLabsAI]

To scale research and deployment of multimodal AGI, today we are pleased to announce that Luma has raised a 900M Series-C and we are partnering with Humain to build a colossal 2GW compute supercluster – Project Halo.

[amit @ CVPR @gravicle]

Holy shit! Ray3 is just a hair below Veo3 and climbing!
SOTA multimodal model from from a co 1/1000th the size of @Google

[Luma @LumaLabsAI]

This is Ray3. The world’s first reasoning video model, and the first to generate studio-grade HDR. Now with an all-new Draft Mode for rapid iteration in creative workflows, and state of the art physics and consistency. Available now for free in Dream Machine.

[Luma @LumaLabsAI]

Introducing Modify Video. Reimagine any video. Shoot it in post with director-grade control over style, character, and setting. Restyle expressive performances, swap entire worlds, or redesign the frame to your vision. Shoot once. Shape infinitely.

[Luma @LumaLabsAI]

Reframe is here. Outpaint and resize any uploaded video, image, or #DreamMachine creation into every format you need. Customize and reposition freely on the canvas to make every version of your story fit perfectly. Goodbye cropping. Hello creative freedom.

[Luma @LumaLabsAI]

You node what time it is. Luma Photon and Ray models are now natively available in @ComfyUI API nodes. Build your wildest workflows now.