AI video generation has leveled up over the last month with OpenAI’s Sora 2 and Google’s Veo 3.1 delivering cinematic fidelity, in-depth narrative control, synchronized audio and massively improved physical realism. Both have significantly upped the ante for creative storytelling.
After testing both separately with a series of wild and imaginative prompts, I had to see how the two most popular AI video generators stacked up against each other. Here’s what happened when I tried some of the craziest ideas I could think of in a faceoff between Sora 2 and Veo 3.1 in a series of the same seven story prompts.
1. The day gravity quit
Prompt: “A sleepy small town on a sunny morning — mailboxes, cars, and coffee mugs start floating upward as gravity gradually turns off. People calmly sip coffee as they rise into the air. The camera tilts skyward, following a golden retriever chasing its leash into the clouds. Cinematic orchestral score, realistic lighting.”
Sora 2 generated a video with depth and detail of a real movie. The soundtrack was especially fitting.
Veo 3.1 delivered a pretty good video, but I was disappointed by typical AI issues like repeating the dog and the flaw of the leash. You’d think by now, those types of inconsistencies would be cleared up.
Winner: Sora 2 wins for greater accuracy and superior storytelling.
2. Grandma’s time-traveling microwave
Watch On
Prompt: “In a cozy 1970s kitchen, a grandma in cat-eye glasses places soup in a microwave that opens a glowing wormhole. Each ‘ding’ flashes a new decade — punk 1980s, cyberpunk 2090, prehistoric cave fire — all visible through the window reflection. Whimsical tone, oversaturated film look, gentle zoom-ins.”
Sora 2 followed through with the prompt and took viewers on a wild adventure with Grandma’s microwave.
Veo 3.1 gave us a quirky grandma with red glasses, which was fun considering the video’s theme, but there weren’t three distinct microwave scenes; but two. Otherwise, I think this one was better visually.
Winner: Sora 2 wins for prompt accuracy.
3. Cloud city jazz club
Prompt: “A floating art-deco jazz club drifts through clouds at sunset. A saxophone player made of shimmering vapor plays for transparent ghost patrons in flapper dresses. The camera cranes down from above the clouds into the lounge. Golden-hour lighting, soft focus, record-crackle soundtrack.”
Sora 2 left me wanting more. The saxophone player seemed to be playing to a dead audience and the overall scene felt dull.
Veo 3.1 felt more cinematic with the wide-angle and zoom, despite the shorter video. The stronger vapor of the saxophone player, combined with the darker colors and livelier audience, made the video feel more like a jazz club.
Winner: Veo 3.1 wins for a more believable and realistic story despite the imaginative prompt.
4. The library at the end of the universe
Watch On
Prompt: “A massive, endless cosmic library — planets orbiting between bookshelves, black holes used as reading lamps. A child floats between shelves in zero-gravity, turning glowing pages that project memories into space. Drone-style camera movement, ambient synth score, volumetric lighting.”
Sora 2 was not as “cosmic” and the books were more like a brick wall.
Veo 3.1 let me down again with the very AI-ish mistake, giving the girl an extra arm and hand. Otherwise, it would have won for a more visually stunning video.
Winner: Sora 2 wins for accuracy, but otherwise, Veo 3.1 would have taken the crown here.
5. Dreams of a broken toaster
Watch On
Prompt: “A retro toaster sits in a kitchen at night, dreaming. In the dream, it imagines itself as a rocket blasting off through a Milky Way made of crumbs and butter pats. The camera follows it like a space documentary. Quirky tone, Pixar-esque realism, twinkly music box score.”
Sora 2 generated an actual story within the video, which felt very much like something small children would watch.
Veo 3.1 created a less interesting video that was too fast and hyper. The breakfast food was odd.
Winner: Sora 2 wins for more accurately following the prompt.
6. Dinosaur news broadcast, 65 Million B.C.
Watch On
Prompt: “A velociraptor news anchor reads headlines behind a stone desk as asteroids streak across the sky behind him. The camera cuts between the anchor, the weather dino, and the live pterosaur traffic report. Cretaceous CNN-style graphics, comedic pacing, realistic feather textures.”
Sora 2 created a very realistic video of a news station with the dinosaur newscasters talking to each other.
Veo 3.1 added extra features, making the storytelling really pop. With more realistic asteroids and unique-looking dinosaurs, this video was more interesting.
Winner: Veo 3.1 wins for superior storytelling and cinematic effects.
7. Humanity’s last disco on the moon
Watch On
Prompt: “A glittering glass dome nightclub on the lunar surface. Astronauts dance in slow motion as Earth rises in the background. The DJ, a humanoid robot with mirrored skin, spins vinyl that floats in zero gravity. The camera orbits 360° around the crowd, strobe lights flashing, 1970s funk soundtrack.”
Sora 2 delivered a rave-like scene with astronauts all dancing uniquely and with personality.
Veo 3.1 embraced the 1970s funk soundtrack with astronauts dancing in unison. The video was visually appealing, and the music felt more accurate than Sora 2.
Winner: Veo 3.1 wins for superior storytelling.
Overall winner: Sora 2
While Veo 3.1 shines in visual cinematic style and mood, Sora 2 consistently delivered more accurate interpretations of complex, surreal prompts — and in many cases, created videos that felt like fully formed stories.
Ultimately, Sora 2 came out ahead in this faceoff, edging out Veo with stronger narrative logic, better object consistency, and smoother emotional tone.
Both Sora 2 and Veo 3.1 represent a leap forward in AI video generation, pushing the boundaries of what’s possible with just a prompt.
Have you tried them? Which one is your favorite? Let me know in the comments.
Follow Tom’s Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button!
More from Tom’s Guide
Back to Laptops
