Wan v2.6 stills are 3 credits at 4K, not the 2.5 T2I row

Wan 2.6 Text to Image is 3 credits at 4K; the sibling Wan Preview 2.5 Text to Image row is 5 credits at 4K — two still pages, two rates, neither is a video generate.

Wan publishes two live stills slugs next to a pile of motion rows. Teams treat them as one “Wan image” button and then argue about a two-credit gap they never named. The gap is the point. wan-v2-6-text-to-image is 3 credits, 4K, discounted on the catalog step. wan-video-2-5-t2i is 5 credits, 4K, not discounted. Both are text-to-image. Both have audio null. Both have no duration. Neither one is a clip.

Two still pages, two rates

Open the catalog by slug, not by family nickname. “Wan stills” is not a SKU. The 2.6 stills page is a generate from text at 3 credits. The 2.5 T2I page is Wan 2.5 in still-image mode — the video family’s frame generator used standalone — at 5 credits. Same ceiling. Different bill. Different composition job.

Text to image is the studio door when you have not named the row. Name the slug. Cost is shown before you confirm. Picking “whatever Wan” is how a 3-credit plate becomes a 5-credit plate without anyone noticing the composition language changed.

Wan’s provider roster is the rest of the family, including motion. This page is only the two stills rates. Do not collapse the two URLs in a brief. wan-v2-6-text-to-image is not wan-video-2-5-t2i. A dispatcher that writes “Wan T2I, 4K” has not chosen.

Wan v2.6 is 3 credits at 4K

Wan 2.6 Text to Image is Wan’s current stills generate: content type image, category text-to-image, requires_image false, max output 4K, listed 3 credits, marked discounted (2 credits on the discounted catalog step). Description on the record: “Text-to-image generation with WAN V2.6.” Aspects listed: 1:1, 4:3, 3:2. Styles: Realistic, Artistic. Qualities: HD, 4K. No 9:16. No 16:9. No duration. Audio does not apply.

You are buying a still you can hold up in a review. Lock ratio on this row, not in a video prompt. If the pack has to ship 9:16 later, that is a different still on a row that lists 9:16, or a crop you accept before anyone spends a video credit. Video models will happily invent a new face at 9:16. That is the failure this slug exists to prevent.

Because this SKU does not require an image, it is a generate, not an edit. If you already have a locked still and need a delta, that is an edit row — the AI photo editor is the surface for instruction-based retouch. Do not re-roll 2.6 and call the new still the same pack member.

3 credits is the listed stills bill. Use it when the job is a 4K plate from text and you do not need Wan-video-native framing. Do not write a camera move into a stills prompt. There is no 5s on this record. There is a 4K image.

The 2.5 T2I row is 5 credits at 4K

Wan Preview 2.5 Text to Image is the other stills page: Wan 2.5 in still-image mode — the video family’s frame generator used standalone. It produces cinematic, video-native compositions that match Wan footage, for first-frame and thumbnail work. Category is text-to-image. 5 credits. 4K. audio is null. No duration. requires_image is false. Not discounted.

This is not a generic poster model that happens to sit next to Wan. The catalog description is specific: video-native compositions that match Wan footage. If the next credit is Wan V2.6 Text to Video — native audio, 5s / 10s / 15s, 1080p, listed 15 credits — or another Wan motion row, this is the still you want to argue about. Not a still from a different family that Wan will restage. Not the cheaper 2.6 stills slug just because both say 4K.

Aspects on 2.5 T2I: 1:1, 4:3, 3:4, 16:9, 9:16. Five shapes. That is a different aspect list than 2.6 stills. If the cut is 9:16, 2.6 stills is the wrong row unless you are cropping on purpose. If the cut is 3:2, 2.5 T2I does not list it. Pick the page that lists the ratio. Do not “fix it in video.”

Styles are Realistic and Artistic. Pick one. There is no audio because there is no time. Do not prompt foley. Those sentences belong on the video row, because Wan’s motion models can emit native audio. This row cannot.

Five credits buys one 4K-capable still, not a clip, not a twelve-angle pack inside one call. First-frame work is the point. A pack generated inside a 15-second video prompt is how you pay motion rates to discover the label is wrong.

Neither row is a video generate

Content type on both records is image. No supports_durations. Audio is unset. Treating either slug as “Wan, but the first frame” is how a stills bill becomes an argument about why the file will not play.

Wan V2.6 Text to Video is the in-family motion generate from text: 5 / 10 / 15 seconds, 1080p, 15 credits, native audio. That is a different slug, a different meter, a different identity sampler. Generating the pack “inside the video model” means you never got a still anyone signed. You got motion that now has to be reverse-engineered back into a hero frame.

The sequence is boring on purpose:

1. Choose the stills page by job. 2.6 stills at 3 credits for a 4K generate on 1:1 / 4:3 / 3:2. 2.5 T2I at 5 credits when the frame has to match Wan footage and you need 16:9 or 9:16 listed.
2. Pick keepers. Kill the almosts. Edit a keeper on the photo editor if the delta is a retouch, not a re-roll.
3. Animate keepers on an image-to-video or Wan text-to-video row after approval.

Skipping step 2 is how a brand kit becomes “whatever the video model felt like this morning.” Skipping the slug choice is how a 3-credit habit silently becomes a 5-credit habit, or how a video-native thumbnail lands on a row that does not list 9:16.

If the yes never happens, do not “just animate it and see.” Motion hides a weak still for one loop. Then every pause frame shows the weakness. Two still pages. Two rates. 3 credits versus 5 credits, both at 4K. Neither is a generate of time.

FAQ

Is Wan 2.6 Text to Image a cheaper video model?

No. Content type is image. Category is text-to-image. 4K, listed 3 credits, no durations. Animate after you approve the still. Wan V2.6 Text to Video is the in-family video generate if the job really is a scene from text.

Why would I pay 5 credits on the 2.5 T2I row?

Because that row is Wan 2.5 in still-image mode: video-native compositions that match Wan footage, with 16:9 and 9:16 listed. If the motion pass is Wan and the pause frame has to belong next to that clip, the extra two credits are the family resemblance. If the motion pass is not Wan, that resemblance is not a reason to be there.

Can I attach a reference photo to either stills slug?

Both list requires_image false and category text-to-image. They are generate-from-text. An edit or image-to-image pass is a different row. Re-rolling either stills slug is a new still, not a retouch. Use the AI photo editor when the plate is already approved and you owe a delta.

Does discounted on 2.6 stills make it the same price as 2.5 T2I?

No. 2.6 stills lists 3 credits (2 on the discounted catalog step). 2.5 T2I lists 5 credits, not discounted. Two pages. Two rates. Name the slug before you confirm.