
Grok Imagine 2.0New
GrokImageBest-in-class instruction following, typography and editing with five references.
354 cr / imageSpecsModel directory21 image9 videoUpdated November 2026
21 image models and 9 video models, from a few dozen credits a render to HD video. Each runs in the same studio on the same credits, so switching models costs nothing but the render. There is no per-model subscription and credits never expire.

Best-in-class instruction following, typography and editing with five references.
354 cr / imageSpecs
The most-liked new open model on Hugging Face. Photoreal in under a second.
54 cr / imageSpecs
Attach up to three pictures and say what to change. Keeps faces, products and text intact.
216 cr / imageSpecs
The full Flux model: strong prompt adherence at a low price.
130 cr / imageSpecs
The December 2025 open Qwen Image. Strong text, more natural people.
216 cr / imageSpecs
Photographic output with less of the AI look.
216 cr / imageSpecs
Krea's open-weight turbo model. The natural Krea look at draft speed.
162 cr / imageSpecs
Edit one picture by instruction. Very good at keeping everything you did not mention.
270 cr / imageSpecs
Tencent's open model built for native 2K output.
270 cr / imageSpecsSharp stills up to 2K with reference editing and multi-image composition.
142 cr / imageSpecs
The cheapest Flux by far. Built for iterating before you commit.
32 cr / imageSpecs
A 17B open foundation model with a painterly, highly detailed default.
259 cr / imageSpecs
Excellent with text inside the image, in English and Chinese.
216 cr / imageSpecs
The fast Flux 2 editor. Up to three reference pictures, the lowest price of the editors.
173 cr / imageSpecs
The slower, higher-fidelity Grok tier for final renders.
494 cr / imageSpecs
The full Z-Image base model: more variety and control than Turbo.
108 cr / imageSpecs
Trained only on licensed data. Built for commercial work that needs a clean provenance.
432 cr / imageSpecs
The fast HiDream model. Dense, painterly detail at a draft price.
130 cr / imageSpecs
Open Flux LoRA, routed through Hugging Face.
177 cr / imageSpecs
The highest-rated open Flux LoRA on the Hub.
177 cr / imageSpecs
Open Flux LoRA with strong contrast and detail.
177 cr / imageSpecsText-to-video and image-to-video up to 720p.
494 cr / secSpecsAttach a still and it moves, with sound. The best value video model in the studio.
173 cr / secSpecsThe open-weight Wan 2.2, routed through Hugging Face. The prompt decides everything.
648 cr / secSpecsTencent's open text-to-video model. Strong prompt following and natural motion.
432 cr / secSpecs
Open image-to-video. Attach a still and it moves.
494 cr / secSpecs1080p image-to-video from xAI.
494 cr / secSpecsThe small Wan 2.2. A five second clip for the price of a few images.
108 cr / secSpecsThe earlier Wan release, with clips up to ten seconds.
648 cr / secSpecsAnimate a still for up to ten seconds.
648 cr / secSpecsThe question is usually not which model is best but which of two is right for one job. These go through that choice on price, quality and what each is for.