There's been some pretty crazy advances in recent weeks in the AI video generation world.
Midjourney has just released their version of video today, and honestly, I am quite impressed considering it's available as part of their regular offering and the motion and prompt adherence is better than their image prompt adherence. I often use Midjourney images for my AI videos to begin with (sent to Kling for example) to have it built it. It has extend option built in. They were trailing hard recently after the release of Sora for images, and v7 was a rushed launch. If looking for an all in one image / video generator with good quality it's a great option. So far, it may replace Kling for me for some video when I want to do quick videos without spending credits. I used it for my latest TikTok video.
Google Veo 3 is also insanely good and commercial level quality with audio included. Main issue is pricing and mostly restricted to Google Gemini Ultra subscription.
Hailuoai Minimax also released 2.0 yesterday. Their new subscription tiers has caused a lot of controversy, as their unlimited plan is no longer (and more expensive as well). It's very good especially compared to 1, but honestly Minimax has fallen behind a bit even with this new version.
A new contender also showed up mysteriously on the AI video battle arena, and it turns out it is Bytedance Seeddance 1.0. It's not out officially to the public yet but it's above even Veo 3 in terms of preferred output.
https://seed.bytedance.com/en/seedance
A year ago we were here At the time, I thought we are already progressing so fast in video, and we have since gone quite beyond