LTX-2.3 is here. For decades, creative software has been defined by its interface. We think the next era gets defined by the engine underneath. LTX-2.3 is a major engine upgrade: → Sharper detail → Stronger motion → Cleaner audio → Native vertical format

A production-ready multimodal engine designed to be built on. And for the first time, we're shipping what we built on top of it too. Any ideas for what’s coming today? Drop in the replies.

API available in one hour. Learn more here:

Big news thank you for open source

Open source always!

You never stop delivering 🤜🤛

Can't stop. Won't stop 💪

Congratulations, good update, hopefully the next update will target bigger model and 96GB of VRAM, I understand there will be a lot of people screaming no one have RTX6000 and it is very expensive, but to reach Seedance V2 level you guys have to push the boundaries, the community will find a way to bring the size down and ComfyUI cloud and the API should be enough to cover everyone.

Improved audio to video?

The time has come to run stuff locally 💪 I’d appreciate ANY advice here 🙏

let's go !!

@indestwas 💪

Great job guys, Ltx on top

Can't wait to try it out! Also, thank you for including my video in the launch video!

Keep creating! 🫡

Bye bye Wan... Hellooooo LTX 2.3

👀

Feels like LTX-2 just came out and we already have 2.3. Local capabilities are skyrocketing!

🚀

Excellent. I didn't know this was coming. I look forward to trying it out.

Can't wait to hear what you think!

native vertical format in a video AI engine update is the most practical shipping decision of the week. every creator distributes vertically now. the tools that treat it as an afterthought lose the workflow. LTX-2.3 built it in.

You get it! 😎

Thank You for this release, can't wait to try it out.

Great work guys.. Thank you!

90s kids didn't even read the text. Our brains went straight here:

Perfect! How and when can we access to the weights?

also 1 more really importand now I remember are you guys gona update for it?

@huggingface ✨👀✨

Congrats, @ltx_model team 👍🏻

My 3090s are ready 🎨

🎉👏🍻

Native vertical support is the bridge between experimental AI video and actual commercial utility for mobile-first creators. By prioritizing the engine over the interface, LTX-2.3 recognizes that the end-product matters more than the tool used to build it. This update specifically solves the friction of cropping and resizing that usually degrades quality. Will we see a surge in high-fidelity AI content on platforms like TikTok now that the resolution matches the consumption?

I don't think my astral 5090 32 vram not enought model its self looks 47 gb

Oh, wow! What an awesome surprise 👏 Can't wait to play with it

@ComfyUI But without good examples for prompt structure for image2video, the result is just a horrible mess. Getting ok at text2image, but image2video is a very different beast...

Klingt spannend! Nutze oft die neuen Features✨, die Ergebnisse sind echt stark. Lass uns austauschen!

This is cool 😎

ltx desktop opensource, but it requires you to input api? can there be any options to use local files ?

"next era gets defined by the engine" is the line. ui is now packaging. who ships the boring workflow first?

Exciting times ahead! Love the focus on what powers creativity. Can’t wait to see what amazing projects come from this upgrade!

Cant wait to see how crisp talking head audio can be out-of-the-box

Working with API. Well done!

🙌🙌

Sharper visuals + stronger motion? This is huge 🔥