Ideogram 4.5 output size per aspect ratio at 1K and 2K on Sume
Exact Ideogram 4.5 pixel sizes on Sume for all 15 ratios at 1K and 2K, plus the four ratios where a 1K edit is larger than text-to-image.
On Sume, Ideogram 4.5 returns a fixed pixel size for each of its 15 aspect ratios: 1024x1024 for 1:1 at 1K and 2048x2048 at 2K, 2560x1440 for 16:9 at 2K. Only 1:3 and 3:1 ignore the tier, and four ratios are larger on a 1K edit than on text-to-image.
The table comes from the size map in the Sume image adapter on main, checked on 2026-10-05 by calling the size resolver directly. No image was generated to produce it.
Sizes by ratio
Send aspect_ratio and resolution (1K or 2K) on POST /v1/images with model: ideogram/ideogram-v4.5. Without resolution the tier is 1K.
| Ratio | 1K | 2K |
|---|---|---|
| 1:1 | 1024x1024 | 2048x2048 |
| 4:3 | 1152x864 | 2304x1728 |
| 3:4 | 864x1152 | 1728x2304 |
| 3:2 | 1248x832 | 2496x1664 |
| 2:3 | 832x1248 | 1664x2496 |
| 5:4 | 1120x896 | 2240x1792 |
| 4:5 | 896x1120 | 1792x2240 |
| 16:9 | 1280x720 | 2560x1440 |
| 9:16 | 720x1280 | 1440x2560 |
| 16:10 | 1280x800 | 2560x1600 |
| 10:16 | 800x1280 | 1600x2560 |
| 2:1 | 1440x720 | 2880x1440 |
| 1:2 | 720x1440 | 1440x2880 |
| 3:1 | 3072x1024 | 3072x1024 |
| 1:3 | 1024x3072 | 1024x3072 |
Edits are larger at 1K for four ratios
When you send input_references, the first image is the one being edited. For that case at 1K, Sume uses a bigger size for four ratios: 16:9 gives 1536x864, 9:16 gives 864x1536, 2:1 gives 1408x704 and 1:2 gives 704x1408. Every other ratio uses the table above, and 2K edits use the 2K column.
So the same aspect_ratio: "16:9" request is 1280x720 as a text prompt and 1536x864 as an edit. Plan your crop or your CSS box from the call type, not from the ratio alone.
What the 2K tier does not do
3:1 and 1:3 are the same at both tiers so the 2K tier adds nothing there. If you need more pixels than that, upscale afterwards instead of expecting resolution: 2K to help.
The size does not change the price. The Fal list price is by quality only (read 2026-10-05), and Sume bills list x 1.25, so picking 2K over 1K costs nothing extra on Sume.
Request
The edit below asks for 16:9 and gets 1536x864 back at the default 1K tier.
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": "Same product, same label, softer window light",
"input_references": [{"image_url": {"url": "https://example.com/src.jpg"}}],
"aspect_ratio": "16:9",
},
timeout=120,
)
r.raise_for_status()
body = r.json()
print(r.status_code, body.get("data", [{}])[0].get("url"))When not to use this table
If you need the output to match your source pixel for pixel, leave aspect_ratio out. An edit with no aspect and no size keeps the source geometry, which is also how the Ideogram API describes its own edit endpoint (read 2026-10-05). Use the table only when you want a new shape, and check the returned image size in code before you publish.
Sources
Related posts
More in Models
- Ideogram 4.5 has two endpoints; on Sume it is one model id
Ideogram lists a precise edit endpoint and a generate-plus-edit endpoint. On Sume one id, ideogram/ideogram-v4.5, covers generate and edit by what you send.
- Ideogram 4.0 has no references; 4.5 takes five. What Sume lists
Ideogram's docs say 4.0 and 3.0 accept no reference images and 4.5 accepts five. Sume lists Ideogram 4.5 and 3; how its edit and reference call works.
- Image launches this week vs Sume's catalog: FLUX 3, Ideogram 4.5
GPT Image 2.5 (Sep 22), Seedream 5.0 Flash (Sep 23), Ideogram 4.5 (Sep 30), FLUX 3 Image (Oct 1): Sume lists GPT 2.5, Ideogram 4.5 and Seedream 5 Lite.
- <IMAGE_REF_0> and <VIDEO_REF_0>: tag references in Gemini Omni
On gemini-omni-flash-1.1 you point at each reference by token in the prompt: <IMAGE_REF_0>, <VIDEO_REF_0>, numbered from 0 in list order. Request example.
Written by Sume