Edit models Morphic names versus what the Sume catalog lists
Morphic runs image edits on Seedream 5.0 Pro, Nano Banana 2 and GPT Image 2.5 until Ideogram 4.5 arrives. Which of these Sume lists, and where to check.

Morphic says Ideogram 4.5 is coming to its canvas and that, until then, image editing there runs on Seedream 5.0 Pro, Nano Banana 2 and GPT Image 2.5. On Sume, GPT Image 2.5 and Ideogram 4.5 are both listed; Nano Banana 2 is retired and its ids run as Nano Banana 2.1; and the docs and pricing code show no Seedream 5.0 Pro, only Seedream 4.5 in the docs and a Seedream 5 lite in the code. Read the live catalog before you plan around any of them.
The comparison
| Morphic names | On Sume | Note |
|---|---|---|
| Ideogram 4.5 | ideogram/ideogram-v4.5 | Listed; edits with references; low, medium, high |
| GPT Image 2.5 | openai/gpt-image-2.5 | Listed; up to 16 references; mask_url |
| Nano Banana 2 | Retired; ids run as Nano Banana 2.1 | Docs: google/nano-banana-2 runs as 2.1 |
| Seedream 5.0 Pro | Not found in docs or pricing code | Seedream 4.5 is in the docs; Seedream 5 lite appears in code |
What the Morphic page does and does not show
The page is a launch explainer for Ideogram 4.5, and it says plainly that its before-and-after examples were not made with Ideogram 4.5. They were made with the models live on Morphic, each source image generated first and then edited with the prompt shown. So the prompts on the page, such as "Change the vase glaze to cobalt blue, keep the light and shadows", are illustrations of the kind of edit, not output from the new model.
That matters for your testing. The claims about untouched pixels, no drift and masks come from Ideogram's description of 4.5. They are not shown on the page. Test them yourself on the model Sume lists, with the script in measuring softening and colour creep.
How to use this list
Be careful with names. Vendors rebrand and version models quickly, and "Pro" or "lite" suffixes mean different capability and price tiers. A model with a similar name in the Sume catalog is not the same model, so match on the id and the description in the list response, not on the marketing name.
If you read a vendor page and want to know what you can call, check three things in order. First, the model id: does GET /v1/images/models list it? Second, its parameters: does it list input_references, and what is the max? Third, price: the pricing array on the endpoint record. The first answer for a retired id is an alias, as with Nano Banana 2, and the job stores the 2.1 id, so your logs show what actually ran.
A Morphic plan is not an API. Morphic lists plans by monthly credits; Sume bills the Image API per image or per token with a documented rate card. Do not compare them without converting to a cost per finished edit.
For a side-by-side of edit features across the Sume catalog, see which Sume model to use for a photo edit and a Python check of references and masks.
Sources
Related posts
More in Comparisons
- Flux 2 Pro to FLUX 3 Image at 1K: 28% higher after Oct 8
Flux 2 Pro bills $0.0375 per image on Sume. A FLUX 3 Image 1K render lists at $0.048 on fal after October 8, 28% more; $0.024 during the promotion, 36% less.
- AI music cost per track: Sume Music vs Lyria 3.5 and ACE Step
Sume Music is a flat $0.125 per generation ($12.50 per 100). Google lists Lyria 3.5 at $0.08 per song ($8.00); fal lists ACE Step at $0.0002 per second.
- Nano Banana 2.1 costs the same as Lite at 1K ($0.0336): what differs
At 1K, Nano Banana 2.1 and Nano Banana 2 Lite both list at $0.0336 per image. The gap is in input and thinking tokens, plus what Lite cannot do.
- Nano Banana 2.1 mask editing: Sume's mask_url is GPT-only
Google's Nano Banana 2.1 adds mask-based editing. On Sume, the mask_url field is accepted only by ChatGPT Image 2.5, so pick the model by the edit you need.
Written by Sume