China's MiniMax launches Sora Competitor

Chinese AI startup MiniMax has unveiled Video-01, its first AI model capable of generating high-resolution videos from text prompts. This is an extremely well funded AI Gen AI startup out of China, see the video demo here.

For a country starved of foreign investments I guess China is doing alright, and they are notably strong in text-to-video solutions and companies having a go at this important domain of text-to-video models.

It is already making waves for its ability to generate hyper-realistic footage of humans, including accurate hand movements. Video-01 is the first iteration of the firm’s Sora-like tool, with future updates expected to enable users to generate videos from images. MiniMax is among China's AI "Tigers", notable startups in Generative AI. The others include  Zhipu AI, Baichuan and Moonshot AI. I have a guest piece coming up on AI Supremacy about the top Chinese AI startups by someone that works at Baidu.

So far it's pretty basic. Video-01 enables a user to input a text description to create a video that is up to six seconds in length. The process from the text prompt to generating a video takes about two minutes. The founder of Minimax is Yan Junjie.

It's not clear to me if MiniMax is going B2C or more B2B and I'm not sure they know yet either. Founded in December 2021, MiniMax offers its new text-to-video-generating tool as part of its consumer-facing Hailuo AI platform, which also provides AI text and music-generating features.

There are so many competitors to Sora already including Runway, Dream Machine, Kling - with several out of China in a fairly surprising move to American researchers and ML scientists. China of course understands how important the next era of video might be - for entertainment, culture and digital companies.

MiniMax founder Yan Junjie stated during a media group interview, ‘We have indeed made significant progress in video model generation, and based on internal evaluations and scores, our performance is better than that of Runway in generating videos.’

• Users can try Video-01 for free at hailuoai.com/video after registering with a mobile number, which also works for non-Chinese numbers.
• ByteDance also has a new Jimeng AI app, text-to-video app available for download on Apple’s mainland App Store, following its July 31 release on various local Android app stores. The desktop version was launched in May.

What's clear is Chinese LLM builders are very interested in creating useful applications for consumers furthering ByteDance's dominance in consumer AI app creation. Curiously, Jimeng allows a user to create 80 images or 26 videos for free and charges a monthly subscription that costs at least 69 yuan (US$9.70) to generate more videos.

MiniMax is such a new startup, so it seems to be currently focused on making the technology widely available rather than commercializing it. You can search MiniMax on X here.

MiniMax has raised a lot and  HongShan, formerly Sequoia China is all in on this startup. To date, MiniMax raised at least $600 million in a funding round led by Alibaba Group Holding Ltd. This round valued the company at more than $2.5 billion. If you know China, it's fairly unusual when a startup has significant funding from both Alibaba and Tencent, which appears to be the case here.

I consider China generally better at creating actual applications for its larger mobile consumer base. With QR code payments and MiniApps, China has historically shown a more innovative mobile digital culture with more convenience and utility for consumers.

Alibaba has itself tried to pick up some of the slack of departing mostly American foreign investors. They have invested in MiniMax, Baichuan Intelligent Technology, Zhipu AI and 01.AI. These are four of the Big Six I am watching mostly intently.