Ideogram 4.5 is coming to Recraft Studio: the API route for edits
Recraft says Ideogram 4.5 is coming to Recraft Studio for drift-free edits. If you need it in code today, Sume lists ideogram/ideogram-v4.5 with edit support.

If you want Ideogram 4.5 for multi-turn edits from code rather than from a design app, call ideogram/ideogram-v4.5 on Sume's image endpoint: send input_references and it edits the first image instead of generating from text. Recraft announced on October 1, 2026 that Ideogram 4.5 is coming to Recraft Studio, alongside Recraft V4.1 and its editing suite, but the post describes a future arrival in that app, not an API.
That is the whole difference. Recraft Studio is where a person drags a photo in and iterates by hand. The Sume Image API is where a script, an agent or a back-office job does the same edit a few thousand times. This post covers what Recraft actually said, what Sume lists for the model today, and how to run a first edit.
What did Recraft say about Ideogram 4.5?
In its October 1 post (read 2026-10-04), Recraft frames the problem as drift: image models change pixels, colors and textures a little more with every pass. It describes Ideogram 4.5 as built for multi-turn editing that keeps everything else in place, and says it edits high-resolution images at their native size, with a demonstration on a 24.2 MP source so edits rejoin the original without visible seams.
Recraft's own wording on availability is that Ideogram 4.5 will sit next to Recraft V4.1 in Recraft Studio, positioned for careful edits rather than first-pass generation. Nothing on that page gives an API route, so treat the Studio integration as a UI feature for now.
What does Sume list for Ideogram 4.5 today?
The facts below come from the Sume Image API page and the repo catalog, read 2026-10-04.
| Item | What Sume lists |
|---|---|
| Model id | ideogram/ideogram-v4.5 |
| Text to image | Send no input_references |
| Edit | First reference is the edited image; up to 4 more are references (5 total) |
| Quality | low, medium, high; omitted means medium |
| List price per image | $0.03, $0.06, $0.22 by quality, whatever the size |
| Resolution | 1K or 2K, combined with aspect_ratio |
Edit without aspect_ratio | Keeps the source image's shape |
How do I run a first edit?
Reference URLs must be public HTTPS. The call below edits one photo, asks for high quality and leaves aspect_ratio out so the result keeps the source geometry. Do not send output_format: the docs say the provider selects it and Sume does not accept the field for this model.
A synchronous call waits up to 30 seconds. Branch on the status code: 200 is the image body, 202 is a job envelope you poll through the standard job endpoints.
import os, requests
r = requests.post(
"https://api.sume.com/v1/images",
headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
json={
"model": "ideogram/ideogram-v4.5",
"prompt": "Change the sofa to dark green. Keep everything else identical.",
"quality": "high",
"input_references": [
{"type": "image_url",
"image_url": {"url": "https://example.com/living-room.jpg"}}
],
},
timeout=60,
)
print(r.status_code)
if r.status_code == 200:
print(r.json()["data"][0]["url"], r.json()["usage"]["cost"])What does a ten-edit chain cost?
Sume bills catalog list price times 1.25, so the three tiers bill at $0.04, $0.08 and $0.28 per image. A ten-pass chain at medium is therefore about $0.80, and the same chain at high is $2.80. Run the early passes at low and only the final pass at high if you are exploring.
Read usage.cost on each response rather than trusting this arithmetic; token counts in usage are always zero on this surface.
Studio or API: which one should you pick?
Pick Recraft Studio when a designer wants to look at each edit and decide the next one. Pick the API when the edit is a repeated rule, such as recoloring a product across a catalog or swapping a price in 40 banners. Because Recraft has not published an API route for the new model in that post, do not plan a pipeline around it until it does.
Sources
Related posts
More in Models
- Runway Gen-4 Turbo takes an image only: text-to-video on Sume
Runway's Gen-4 Turbo accepts image input only. If your prompt has no source image, these Sume video models take text alone, and these two need a picture.
- Runway lists Seedance 2.5 with video input: what Sume accepts
Runway says seedance2_5 takes text, image or video. On Sume, seedance-2.5 takes video as a reference, not an edit source. Which model edits a clip.
- Scribe v2, Realtime or Medical: which model for interview clips
ElevenLabs lists Scribe v2 with up to 32 speakers and 90+ languages, Realtime at about 150 ms, and Medical with 35% fewer errors. Which fits interview clips.
- Seedance 2.0, Fast and Mini next to 2.5: still callable on Sume
Sume still serves seedance-2, seedance-2-fast and seedance-2-mini by id next to seedance-2.5. 15-second ceiling, per-token prices, 4 s and 5 s job costs.
Written by Sume