Seedance 1.5 Pro is now live on CapCut app. Seems like this is how ByteDance's AI models go globally. Just unceremoniously on its wide suite of apps.
Alibaba's Wan 2.6 is also out.
Sensetime launches
Artificial Intelligence thread
- Thread starter 9dashline
- Start date
Seedance 1.5 Pro is now live on CapCut app. Seems like this is how ByteDance's AI models go globally. Just unceremoniously on its wide suite of apps.
Alibaba's Wan 2.6 is also out.
Sensetime launches
Do you have the direct link to this AI?
I would like to try out the PPT feature. I've been using notebook lm for it and kimi but they are not great
Xiaomi out with their latest open source LLM. Actually looks pretty good for a first major effort.
Meituan is also out with another model that allows generating video based on audio and text.
Now Alibaba finally announced this Wan 2.6
check here
The Xiaomi MiMo model in particular looks very interesting. I suspect AI will be dominated by big tech companies that can self-fund expensive AI R&D through free cash flow.
Moreover, of all the major AI players in China, Xiaomi is the only one which produces cars. I don't know if they produce their own robots, but it is inevitable that they will probably start to compete with Ubtech and other robotics players head to head. It makes sense to be present in the entire ecosystem end-to-end and few companies are better placed to do so than Xiaomi.
Moreover, of all the major AI players in China, Xiaomi is the only one which produces cars. I don't know if they produce their own robots, but it is inevitable that they will probably start to compete with Ubtech and other robotics players head to head. It makes sense to be present in the entire ecosystem end-to-end and few companies are better placed to do so than Xiaomi.
Xiaomi usually tends to join the game when the market is mature enough, like it used to do in smartphone and EV. The embodied intelligence is far away from that goal.
With Gemini 3.0-Flash, Western closed source labs have finished their latest round of LLM releases, and here is where Artificial Analysis rankings currently stand. It's clear Google has taken the lead with 2 out of the 3 top models, I'm genuinely surprised that Gemini 3.0-Flash is that close to 3.0-Pro while being much faster / cheaper, but still have to use it to see if it is just bench maxed. Significantly, Open AI's "code red" emergency GPT 5.2 release did not over take Google, indicating they are now slightly behind since Google released their model first and forced Open AI into an off-cycle release.
(Not going to bother with LLMArena since it is highly biased to Western tastes / votes).
The top five spots are all occupied by Western models, but the difference between them and the closest Chinese competitors (Kimi / Deep Seek) is only 6 points. I look forward to see Chinese labs' responses, since Chinese labs are about ~2-3 months out of their last releases (except Deep Seek) and so are due for the next cycle.