The ladder, read carefully

The cleverness of this launch is how the two price sheets talk to each other. The flagship's three rungs are $0.08, $0.14 and $0.25. Lite's are $0.02, $0.03 and $0.14. The budget tier's top price is the flagship's middle price, and the flagship's bottom price is where Lite's ladder ends entirely. There is no overlap, which is the point: the tiers are a staircase with no landing in the middle.

Now do the per-clip math, because per-second pricing flatters short clips and punishes everything else. A default six-second generation at 480p costs $0.12, plus a penny if you steer with an image. The same clip at the “1080p” tier costs $0.84, for pixels rendered at 720p. At a full 15 seconds, the upscaled 1080p tier runs $2.10. Volume work at the top tier stops looking like a bargain fast.

Two cents buys drafting, not delivery

None of this makes Lite bad. It makes it specific. At $0.02 to $0.03 a second with native audio, it is the cheapest way to run a lot of generations while you figure out what works: motion tests, prompt iteration, client previews, the hundred clips that never ship. That is the job the 480p and 720p tiers are priced for.

The things Lite leaves out say as much. The flagship adds multi-reference control (up to seven image references) and voice references for holding a character's face and voice across scenes. If your clip needs the same person in the next clip, Lite is not the tool. Budget pricing with no consistency controls is a familiar formula in this market: it is how you price a model for experimentation without cannibalizing the production tier.

What it means for the price war

Kling's 3.0 Turbo charges $0.11 a second for 720p with audio. Lite charges $0.03. That gap is the story at the low end, and it is real money at volume: a thousand six-second 720p test clips run $180 on Lite versus $660 on Kling. At 1080p the two converge at $0.14, and there Lite's advantage is just distribution (Venice's per-clip pricing, OpenRouter routing, xAI's own API) rather than price.

The read on xAI's move is straightforward. Video generation is becoming infrastructure, and infrastructure competes on price per unit. xAI is not selling a better model here. It is selling a cheaper slot in the pipeline, with the expensive slots kept exactly where they were.

Sources

  1. [1] xAI API pricing docs — “API Pricing”Read source
  2. [2] OpenRouter — “Grok Imagine Video 1.5 Lite - API Pricing & Benchmarks”Read source
  3. [3] Zeniteq — “xAI Adds Grok Imagine Video 1.5 and a $0.02 Lite Tier”Read source
  4. [4] sloptvnews — “xAI's Grok Imagine Video 1.5 Lite ships from $0.02 a second”Read source
  5. [5] Venice.ai — “Grok Imagine 1.5 Lite”Read source