Most image-to-video models take one starting photo and imagine where it goes next. The 3 models below take two images — a first frame and a last frame — and generate the motion that connects them, which is a more precise brief than plain image-to-video and the only way to guarantee a specific end state.
This is a thin field by catalog standards: only 3 providers have built it (Wan, Black Forest Labs, Google), and each shipped it independently rather than converging on one tag name — the catalog calls it "first_last_frame" on some models and "frame control" on others.
Ranking method
Filtered to models whose feature tags name first-and-last-frame or frame control, sorted by best leaderboard position (unranked models after, cheapest complete job first).
Full ranking
| # | Model | Provider | Credits | Max resolution |
|---|---|---|---|---|
| 1 | Wan 2.7 Image to Video | Wan | 16-180 credits | 1080p |
| 2 | Flux 3 First Last Frame to Video | Black Forest Labs | 34-136 credits | 1080p |
| 3 | VEO First Last Frame | 64-256 credits | 4K |
The top 3, explained
#1 — Wan 2.7 Image to Video (Wan) costs 16-180 credits and sits at #22 in image to video on Versely's live model rankings. It outputs up to 1080p. Max resolution: 1080p.
#2 — Flux 3 First Last Frame to Video (Black Forest Labs) costs 34-136 credits and does not currently hold a position on any Versely leaderboard. It outputs up to 1080p. Max resolution: 1080p.
#3 — VEO First Last Frame (Google) costs 64-256 credits and does not currently hold a position on any Versely leaderboard. It outputs up to 4K. Max resolution: 4K.
More capability rankings
Frequently asked questions
What is the best AI model for first-and-last-frame video?+
Wan 2.7 Image to Video by Wan tops this ranking. Filtered to models whose feature tags name first-and-last-frame or frame control, sorted by best leaderboard position (unranked models after, cheapest complete job first). Wan 2.7 Image to Video costs 16-180 credits and sits at #22 in image to video on Versely's live model rankings.
How is this ranking calculated?+
Filtered to models whose feature tags name first-and-last-frame or frame control, sorted by best leaderboard position (unranked models after, cheapest complete job first).
How many models qualify for this ranking?+
3 models with a page on Versely meet the criteria for "Best AI model for first-and-last-frame video", across 3 providers.
Is the top-ranked model also the cheapest?+
Yes — Wan 2.7 Image to Video is both the top entry and the cheapest option here, at 16-180 credits.
Do all of these models hold a Versely leaderboard position?+
1 of the 3 models here hold a position on at least one Versely leaderboard; the remaining 2 are unranked and listed afterward, cheapest complete job first.
Try Wan 2.7 Image to Video inside Versely
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync, in your browser or on your phone.
Free on iPhone. On a computer? The same account works on Versely Web, no install needed.
Guides
A first-last-frame video interpolates two stills, not one prompt
generate_video_from_image in first-last-frame mode needs two photos. One still is a different job, and faking the second wastes credits.
Flux 3 First Last Frame to Video: the still is the contract (5s, 6s, 7s, 9cr)
Flux 3 First Last Frame to Video is image-to-video. If the label is wrong, every second is waste.
First/Last Frame: Perfect Transitions and Seamless Loops
First/last frame AI video control explained: seamless loops, invisible transitions between scenes, morph reveals, and frame-pair design rules.
Veo first-last frame for a wine-barrel bung pull
A seated bung and a pulled bung on the same wine barrel interpolate on Veo first-last at 40cr, 4K, 4s / 6s / 8s, native audio; /for/wineries is the cellar page, not this stave SKU.