We have a new #1 AI video generator! (beats Veo 3)

AI SearchAbout 7 min readJun 19, 2025Watch original
THE SUMMARYAI-generated

AI Video Generator: Hyre O2 Model - Detailed Analysis and Comparison

Key Concepts:

  • AI Video Generation
  • Text-to-Video
  • Image-to-Video
  • Prompt Understanding
  • Physics Understanding
  • Camera Control
  • Coherence
  • Deformation
  • Google V3
  • Cling 2.1
  • ELO Score

1. Introduction to Hyre O2

  • Hyre has released its flagship O2 model, an AI video generator.
  • The presenter claims it's the best AI video generator currently available, surpassing even Google's V3 in certain aspects.
  • The video will cover the capabilities of Hyre O2, compare it to Google's V3 and Cling 2.1, and discuss its performance, specs, and usage.

2. Getting Started with Hyre O2

  • Access Hyre's website (link in description).
  • New accounts receive 500 free points.
  • Users can generate videos using the latest Hyre O2 model.
  • Two generation methods are supported:
    • Image-to-Video: Upload an image as the starting frame.
    • Text-to-Video: Create a video from a text prompt.
  • Hyre O2 creates videos with a resolution of 768p.
    • Users can choose between 6-second and 10-second videos.
    • 1080p videos are supported, but only for 6-second durations.

3. Text-to-Video Examples and Comparisons

  • Hyre O2 excels at understanding complex prompts, including camera movements and physics.

    • Example 1: Man Riding a Unicycle and Juggling Red Balls

      • Hyre O2 successfully generated a video of a man riding a unicycle and juggling red balls.
      • Minor flaws were noted with the balls disappearing near the end.
      • Comparison:
        • Google V3: The unicycle doesn't move, and the juggling is incorrect.
        • Cling 2.1: The unicycle moves, but the man isn't pedaling or juggling correctly.
      • Result: Hyre O2 wins this round.
    • Example 2: Will Smith Eating Spaghetti

      • Hyre O2 generated a realistic video of Will Smith eating spaghetti.
      • Other video generators failed to produce a similar result.
      • Google V3: Censorship prevented video generation.
      • Cling 2.1: Generated a video of a white man who doesn't resemble Will Smith.
      • Result: Hyre O2 wins this round.
    • Example 3: A Skier Launches Off a Snow Ramp Twisting Acrobatically in the Air

      • Hyre O2 generated a video of a skier twisting acrobatically in the air without deformation.
      • Comparison:
        • Google V3: The skier's head appears from behind near the end.
        • Cling 2.1: The skier skis backward and performs unrealistic movements.
      • Result: Hyre O2 wins this round.
    • Example 4: A Gymnast Performs a Flip on a Balance Beam

      • Hyre O2 generated a video of a gymnast doing a flip on a balance beam.
      • The flip wasn't perfect, but the gymnast's limbs remained intact.
      • Comparison:
        • Google V3: Warping and deformations occurred, with legs disappearing and reappearing.
        • Cling 2.1: The quality was good, but the flip was incorrect.
      • Result: Hyre O2 wins this round.
    • Example 5: A Giant Cat Wearing Aviator Goggles Races Across a Desert with a Child Clinging on Its Back. The Cat Runs So Fast That It Leaves a Trail of Fire

      • Hyre O2 generated a video of a giant cat wearing aviator goggles racing across a desert with a child clinging on its back, leaving a trail of fire.
      • Comparison:
        • Google V3: Could not generate this.
        • Cling 2.1: The cat's tail was on fire instead of leaving a trail of fire.
      • Result: Hyre O2 wins this round.
    • Example 6: A Group of Friends Having Brunch in a Cozy Cafe. Suddenly, a Capybara Falls on the Table Surprising Everyone

      • Hyre O2 generated a video of a group of friends in a cafe with a capybara falling on the table.
      • The physics were accurate, with the capybara crashing on the plates of food.
      • Comparison:
        • Google V3: The capybara fell off the table, and the food wasn't affected.
        • Cling 2.1: The capybara was already on the table, and everyone ran away.
      • Result: Hyre O2 wins this round.
    • Example 7: A Princess Wearing a Glittery White Dress. She Is Running Away From a Massive Dragon with Glowing Red Eyes. 3D Disney Pixar Style

      • Hyre O2 generated a video of a princess running away from a dragon with glowing red eyes in 3D Disney Pixar style.
      • Comparison:
        • Google V3: The dragon didn't have glowing red eyes and appeared out of nowhere.
        • Cling 2.1: The dragon ended up in front of the princess.
      • Result: Hyre O2 wins this round.
    • Example 8: The Camera Pushes in Towards the Couple Kissing. The Man Is in a Black Suit. The Woman Is Wearing a Red Dress. Then Tilts Up to Show the Sky and Overlay Text The End

      • Hyre O2 generated a video with the specified camera movements, clothing, and text overlay.
      • Comparison:
        • Google V3: Ignored the camera movements.
        • Cling 2.1: Pushed in the camera, but didn't tilt up, and the text was gibberish.
      • Result: Hyre O2 wins this round.
    • Example 9: A Man in a Denim Jacket Sitting at a Small Kitchen Table in a Cluttered New York Apartment. The Table Is Covered in Old Newspapers, an Empty Coffee Mug, and a Half-Eaten Bagel. A Cat Naps on the Windowsill. Through the Window You Can See a Bear. His Roommate Wearing a Bathrobe and Sunglasses Walks In

      • Hyre O2 generated a video with all the specified elements.
      • Comparison:
        • Google V3: Also got it correct.
        • Cling 2.1: Missing the bear.
      • Result: Tie between Hyre O2 and Google V3.
    • Example 10: A Live Broadcast of a FIFA World Cup Soccer Match

      • Hyre O2 generated a video with good clarity and physics.
      • Comparison:
        • Google V3: Deformationations and gibberish text.
        • Cling 2.1: Physically incorrect, with a small field and missing ball.
      • Result: Hyre O2 wins this round.
    • Example 11: A Motorcycle Racer Leaning Sharply into a High-Speed Turn

      • Hyre O2 generated a sharp and coherent video.
      • Comparison:
        • Google V3: Also got it correct.
        • Cling 2.1: Also got it correct.
      • Result: Tie between all three.

4. Image-to-Video Examples and Comparisons

  • Image-to-video allows for greater control over the video's composition.

    • Example 1: Warrior Sprinting Towards a Monster

      • Hyre O2 generated a video of a warrior sprinting towards a monster that breathes fire.
      • The scene was coherent, with no warping or deformation.
      • Comparison:
        • Google V3: Lacked movement and cinematic quality.
        • Cling 2.1: The warrior's limbs deformed badly.
      • Result: Hyre O2 wins this round.
    • Example 2: Orbit Shot of Dudes Having an Intense Fight

      • Hyre O2 generated a video with an orbit shot, a sword clash, and a zoom into the characters' faces.
      • The generation was flawless and followed the prompt well.
      • Comparison:
        • Google V3: Lacked action.
        • Cling 2.1: Turned the swords into lightsabers and didn't follow the camera movements.
      • Result: Hyre O2 wins this round.
    • Example 3: Anime Girl Talking to a Creature

      • Hyre O2 generated a video of an anime girl talking to a creature.
      • Comparison:
        • Google V3: Didn't follow the prompt, and the girl didn't appear to be talking.
        • Cling 2.1: The girl wasn't talking to the creature.
      • Result: Hyre O2 wins this round.
    • Example 4: Panning to Reveal a Girl in a White Dress

      • Hyre O2 generated a video that panned to the right to reveal a girl in a white dress.
      • Comparison:
        • Google V3: Could not generate this.
        • Cling 2.1: Didn't pan, and there was no girl in a white dress.
      • Result: Hyre O2 wins this round.
    • Example 5: Fast Zoom Up to the Girl's Face Showing a Very Scared and Surprised Expression

      • Hyre O2 generated a video with a fast zoom up to the girl's face, showing a scared expression, and followed the prompt perfectly.
      • Comparison:
        • Google V3: Forced to crop the image into a horizontal widescreen ratio. Lacked action.
        • Cling 2.1: Looked really good and followed the prompt exactly.
      • Result: Tie between Hyre O2 and Cling 2.1.

5. Overall Performance and Conclusion

  • Hyre O2 is noticeably better than Google V3 and Cling 2.1 in prompt understanding, camera control, coherence, and physics understanding.
  • According to an independent leaderboard by Artificial Analysis, Hyre O2 beats Google V3 by almost 100 ELO points for image-to-video.
  • Hyre O2 is currently the best video model available, as ByteDance's Cance 1.0 has not been released yet.

6. Pricing

  • A free trial of 500 credits is available.
  • 768p videos cost 25 credits.
  • Full HD videos require a paid plan.
  • The cheapest plan is $96 per year ($8 per month).
  • Google V3 costs $250 per month.

7. Final Thoughts

  • The presenter encourages viewers to try Hyre O2.
  • Viewers are asked to share their experiences and generated videos in the comments.
  • The presenter will continue to share top AI news and tools.
  • Viewers are encouraged to subscribe to the presenter's YouTube channel and newsletter.

AI summaries can miss context or contain errors. Check important details against the original video.

MAKE IT YOURS

Read. Remember. Reuse.

Free tools

Go a little deeper.

Have a question about this video? Load its transcript to open the video chat.