How many aspect ratios each Sume image model lists, vs Hy's 14
Hy Image 3.5 Preview lists 14 aspect ratios. On Sume, Nano Banana 2.1 and Ideogram list 15, Grok and Flux 13, GPT Image 2.5 only 9. Full counts by row.

Tencent's Hy Image 3.5 Preview lists fourteen aspect ratios, according to its OpenRouter page (read 2026-10-08). Sume does not list Hy Image 3.5. Among the image models Sume does list, Nano Banana 2.1 and the Ideogram rows list 15 ratios, Grok Imagine and the Flux, Qwen and Recraft rows list 13, and ChatGPT Image 2.5 lists 9.
Counts by row
The counts come from the per-model aspect lists that back GET /v1/images/models, where each row publishes an aspect_ratio descriptor with its allowed values. Counts include auto where the list has it. The lists differ in shape as well as in length, so a longer list is not always better for your layout.
| Row | Values | Which ones |
|---|---|---|
| Nano Banana 2.1 | 15 | auto, 21:9 down to 9:16, plus 4:1, 1:4, 8:1, 1:8 |
| Ideogram V3 and 4.5 | 15 | 1:3, 3:1, 10:16, 16:10 and 4:5 among them; no auto |
| Grok Imagine | 13 | 2:1, 20:9, 19.5:9, 9:19.5, 9:20, 1:2 among them |
| Flux 2 Pro and Flex, Qwen Image and Max, Recraft V4 | 13 | 1:1, 16:9, 9:16, 4:3, 3:4, 4:5, 5:4, 3:2, 2:3, 21:9, 9:21, 1:2, 2:1 |
| Nano Banana Pro | 11 | auto, 21:9 to 9:16; no strip ratios |
| Seedream v4 | 10 | auto plus the nine Seedream ratios |
| Seedream 5.0 Lite and 4.5 | 9 | 1:1, 16:9, 9:16, 4:3, 3:4, 4:5, 5:4, 3:2, 2:3 |
| ChatGPT Image 2 and 2.5 | 9 | auto, 1:1, 16:9, 9:16, 4:3, 3:4, 5:4, 9:8, 4:5 |
| Higgsfield Soul | 7 | 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3 |
| Imagen 4 Fast and Ultra | 5 | 1:1, 16:9, 9:16, 4:3, 3:4 |
Reading the table
Length is a poor proxy for fit. Ideogram's 15 values include tall ratios such as 1:3 and 10:16 that no Flux or GPT row lists. Nano Banana 2.1 is the only row with the 8:1 and 1:8 strips. GPT Image 2.5 has the shortest list among the edit-capable rows, but it also accepts custom pixels through image_size, so the list is not its ceiling.
Grok Imagine's list is tuned to phone screens, with 19.5:9, 9:19.5 and 9:20, and it has no 4:5. Imagen 4 has five ratios and no custom pixels.
Check a ratio in code
Do not copy this table into your app. The catalog is the source of truth and it changes. Read the descriptor at runtime and fall back to a row that lists your ratio.
import os, requests
H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
want = "9:20"
r = requests.get("https://api.sume.com/v1/images/models", headers=H, timeout=30)
r.raise_for_status()
for m in r.json()["data"]:
ar = m["supported_parameters"].get("aspect_ratio", {})
if want in ar.get("values", []):
print(m["id"], len(ar["values"]))What the Hy page does and does not say
The OpenRouter listing says fourteen options including 1:1, 16:9, 9:16, 4:3 and 3:2. It does not name all fourteen in the text we could read, so we make no claim that a particular unusual ratio is or is not on its list. If you need a ratio from Hy, check the vendor page before you plan around it, and check Sume's descriptor for the row you would use instead.
Sources
Related posts
More in Models
- HunyuanVideo VRAM: 45 GB at 544p, 60 GB at 720p, 80 GB advised
The HunyuanVideo README lists 45 GB for 544x960 and 60 GB for 720x1280 at 129 frames, tested on one 80 GB GPU. FP8 saves about 10 GB. Table and hosted route.
- HunyuanVideo 720p: 1,904 s on one GPU, 338 s on eight
README latency for a 1280x720, 129-frame, 50-step clip: 1,904 s on 1 GPU and 338 s on 8. The GPU-seconds each setup burns, and what it means for hosted jobs.
- Hy Image 3.5 lists a 1.5K tier. No Sume image model does
Hy Image 3.5 Preview offers 1K, 1.5K, 2K and 4K. Sume's tier lists are 512, 1K, 2K, 4K or 1K/2K only. For about 1.5K, GPT rows take 1536x1536 as image_size.
- Hy Image 3.5 Preview: 2K or 4K? Tencent and OpenRouter differ
Tencent's launch post says up to 2K; OpenRouter's page lists resolution 1K to 4K. How to test it, and which Sume image rows list a 4K tier or pixel size.
Written by Sume