Wan 2.5 Takes A Shot at the AI Video Throne!

Another bonkers week in AI video. Alibaba’s WAN 2.5 has officially landed, and this one is a game-changer. Not only does it generate dialog + audio, but it also takes its first real shot at Google’s Veo 3. We’ll walk through side-by-side examples, tests, and community outputs to see how WAN stacks up. Plus:Google Flow updates with new prompt expanders and NanoBanana integrations! Dreamina’s long-form breakthrough: 45s–1 minute seamless generations with up to 10 keyframes, Cost, platforms, and open-source rumors for WAN 2.5 Big Thanks to today's Sponsor, Emergent! Check them out here: https://app.emergent.sh/?via=theoreticallymedia 👉 Try WAN on the official site: https://create.wan.video 👉 Test WAN on Higgs Field: https://higgsfield.ai/ 👉 Google MixBoard experiment: https://labs.google.com/mixboard/welcome If you’re into AI filmmaking, text-to-video, Runway, Veo 3, Sora, or WAN, this breakdown covers it all. 00:00 – Another wild week in AI video 00:18 – WAN 2.5 release: first real Veo 3 challenger 00:44 – Audio + dialog generation explained 01:12 – Platforms, free vs credits, and cost overview 02:05 – First text-to-video test: futuristic monologue 03:40 – Simulation romance test (camera moves & detail) 04:31 – Found footage “Buffy” project test 05:22 – Vampire fight sequence + lip sync issues 05:48 – Background music tests (accidental musicals) 06:11 – Image-to-video: Flamethrower Girl & anomalies 07:08 – Subtitles + focal shift surprise 07:23 – Cultist summoning a fire demon 08:16 – Femme fatale detective office test 09:11 – Jason-inspired run cycle & stutter test 09:48 – Fixing lip sync with sound-driven editing 10:18 – Community experiments: audio-to-video stick figure test 11:12 – YouTube AI Search & Higgs Field demos 12:01 – WAN 2.5 on Korea platform + API hiccups 12:12 – Pricing breakdown: credits vs Veo 3 12:56 – Open-source rumors: preview vs release 13:16 – Why WAN 2.5 matters: dialog competitor 13:22 – Sponsor Segment: Emergent (ad spot) 16:04 – Google responds: Flow platform & MixBoard 16:30 – Flow updates: prompt expander & Viking demo 19:03 – Ingredients-to-video missing but teased 19:34 – 1-minute keyframe generations 20:24 – Long-take seamless transitions tested 21:01 – Wrap-up: next drops ahead #WAN25 #AIvideo #Veo3 #TextToVideo #AItools #Filmmaking