The cheapest AI video models, in order

Seedance 1 Pro text-to-video, accelerated.

The cheapest Luma tier, at 540p.

The Kling 2.5 look, at the standard tier.

H3 Max with the brakes off, at half the cost.

Fast iteration, with a draft mode.

The small, cheap Wan for quick 480p drafts.

Hailuo 2 at the standard tier, a steady low cost option.

Pixverse v6, strong on stylised and animated looks.

Kling 1.6 at 720p and 30fps, the budget entry point.

The most cost efficient way to animate an image.

Seedance 1.5 Pro for text-to-video, with audio.

The lightest Veo 3.1 tier.

xAI's Grok Imagine video model.

The original Kling generation, at the lowest price.

Kling 2.1 image-to-video, a proven older generation.

Hailuo 2.3 at the standard tier, tuned for value.

Seedance 1 Pro for text-to-video.

The previous LTX generation, tuned for speed.

Ray 2's look, faster and cheaper, at 720p.

Seedance 1 Pro, faster and cheaper.

Hailuo 2.3 with lower latency, keeping its motion quality.

Kling 2.5 Turbo Pro, a strong balance of speed and quality.

Kling 2.5 Turbo Pro, professional output at speed.

H3 retrained for closer prompt adherence.

The light Seedance tier, for text and image driven clips.

Accelerated Wan 2.1 text-to-video at 480p.

The dependable all-rounder for everyday production work.

Faster, cheaper, and now takes reference images.

Expressive motion from a still, with sound.

Kling 1.6 at full 1080p.

Kling 2.1 at up to 1080p, in 5 and 10 second clips.

Built for speed, with native audio and high resolutions.

Accelerated Wan 2.1 image-to-video at 480p.

Preview build of Wan 2.5.

Wan 2.7, with first and last frame control.

Hailuo 2, strong on physical realism.

High fidelity model tuned for realistic human motion.

The Kling O3 standard tier.

Kling v3 quality at turbo speed.

Flagship generation with synchronised sound.

The original Hailuo model, in 6 second clips.

Video-01 with explicit camera direction.

Trained for Live2D and illustrated characters.

Wan 2.5 tuned for speed, from an image.

Wan 2.5 tuned for speed, from a text prompt.

Smoother motion and cleaner scenes than Wan 2.2, up to 1080p.

Wan 3.0 driven by reference images rather than one frame.

Google's Gemini Omni Flash, built for speed.

Runway Gen-4.5, strong motion quality and prompt adherence.

Joint audio and video, following complex instruction.

Expressive motion with reference image control.

Alibaba's Happy Horse text-to-video model.

Happy Horse 1.1, refined over the first release.

The Kling O3 professional tier.

Kling 2.6 Pro, cinematic image-to-video with fluid motion.

Kling 2.6 Pro, steadier motion than the 2.5 line.

The professional Kling v3 tier, accelerated.

The professional Seedance 1 tier.

Veo 3 at a lower price and a shorter wait.

Veo 3.1 accelerated, for iterating on a shot quickly.

The flagship Wan tier, for shots that carry the piece.

Coherent, physically plausible motion from a single frame.

Cinematic framing with fluid, believable movement.

Seedance 2.0 at speed.

The smallest Seedance 2.0 tier.

Luma Ray 2 at 720p, in 5 and 9 second clips.

Top tier cinematic image-to-video, with optional audio.

Accelerated Wan 2.1 image-to-video at 720p.

Kling 2.0, a clear step up from the 1.x line.

The premium 2.1 tier, with stronger dynamics.

Seedance 2.0 for text-to-video, with audio.

The current flagship, strongest on realism and natural motion.

Google's flagship Veo 3, with generated audio.

Kling 3.0 Omni, multimodal generation with references.

Kling 3.0, cinematic clips of up to 15 seconds.

Single shot clips up to 30 seconds, with audio.

Seedance 2.5 driven by reference material.

The most capable Sora tier, with sound.

Google Veo 2, faithful to written direction.