ChatGPT now has more Models than Paris Fashion Week

Kling 2.0 is here, and it’s turning heads with its next-level AI video generation. If you thought Sora was impressive, Kling is out to steal the spotlight. Built by Kuaishou (TikTok’s rival in China), Kling 2.0 can generate full-blown HD videos up to 2 minutes long with photorealistic people, smooth camera motion, and—get this—physics-aware movement. It even handles walking, dancing, and realistic body transitions like a champ. People are already calling it the most Sora-like model outside of OpenAI. From city scenes to cinematic close-ups, Kling nails it all with shockingly few input prompts. And the motion quality is clean, fluid, believable. It’s still early days, but if this is where AI video is heading, traditional filming is about to get seriously disrupted. Hollywood, marketers, and indie creators alike should be watching this space very closely.

Canva just rolled out Visual Suite 2.0, and it’s a creative powerhouse. This update brings a wave of new AI tools built to make design faster, smarter, and way more fun. It features Magic Media, which can now generate AI videos and animations right inside Canva. There’s also an upgraded Magic Design that whips up polished presentations or social posts based on just a sentence. Plus, the new AI-powered photo editor can remove objects, change backgrounds, or even swap faces—no Photoshop skills needed. What makes Visual Suite 2.0 stand out is that it’s not just for designers. Anyone, from marketers to small biz owners, can now create brand-level content in minutes. Canva has turned into a full-blown content engine, think presentations, videos, social graphics, printables, all in one flow. If your team’s been juggling five different design tools, this might be your all-in-one fix. Canva’s aiming to be the Photoshop, Premiere, and Figma for everyone.

TikTok’s parent company, ByteDance, just introduced its new AI video model: Seaweed. The name might sound quirky, but the capabilities are anything but that. Seaweed generates 16-second, 1080p videos with smooth motion and realistic scenes. Early demos show a solid grasp of physics walking, object movement, and even some cinematic camera pans. While it’s not as fluid as OpenAI’s Sora yet, it’s a big step forward and a clear signal, ByteDance isn’t playing around in the AI space. What’s interesting is how Seaweed balances creativity and realism. Whether it’s generating stylized animations or real-world scenes, it holds its own. ByteDance is already a giant in short-form video, imagine this tech embedded natively inside TikTok. That’s a whole new era of AI-generated content coming straight to your For You page. No public access yet, but insiders say it’s being tested by creators already.

OpenAI quietly rolled out ChatGPT-4.1 this week alongside two new models—o3 and o4-mini—and the updates are solid. ChatGPT-4.1 brings faster response times, better memory handling, and smoother contextual understanding. It’s noticeably more “aware” in conversation, and that makes it easier to use for complex tasks without repeating yourself or overloading it with reminders. It also appears to have improved reasoning and fewer hallucinations. The o3 and o4-mini models. These are leaner, faster versions built for performance on smaller devices and real-time use cases—think mobile apps, smart assistants, and embedded systems. Early reports say o4-mini runs impressively well even on consumer-grade hardware. These releases are part of OpenAI’s strategy to scale ChatGPT beyond just the web and into the real world. If you’re building anything with AI right now, o3 and o4-mini could be the new go-tos for fast, cost-efficient integration. More speed. Less fluff. And just in time for devs to start testing ahead of GPT-5 rumors.

Tavus is a generative AI video research company that enables developers to build digital twin video experiences through easy-to-use APIs.