ChatGPT Image 2.5 low vs Ideogram 4.5 low: 1 cent vs 4 cents
Low tiers on Sume: ChatGPT Image 2.5 is 1 cent, Ideogram 4.5 is 4. Omit quality and GPT runs at high (7 cents), Ideogram at medium (8).

At the low tier, ChatGPT Image 2.5 is about 1 cent per image on Sume and Ideogram 4.5 is 4 cents. At the defaults, the two are nearly equal: ChatGPT Image 2.5 runs at high when you omit quality (6 to 7 cents), and Ideogram 4.5 runs at medium (8 cents). So the same lazy request costs almost the same on both, while a deliberate quality: "low" request costs four times more on Ideogram.
Both ladders
Sume bills the catalog list times 1.25, rounded up to the cent. Ideogram's list is $0.03, $0.06 and $0.22 per image by quality, as the Sume Image API page documents. The ChatGPT Image 2.5 catalog list of $0.0527 is the high 1024 rate.
| Quality | ChatGPT Image 2.5 | Ideogram 4.5 |
|---|---|---|
| low | 1 cent | 4 cents |
| medium | 2 cents | 8 cents (Ideogram default) |
| high | 6 to 7 cents (GPT default) | 28 cents |
| xhigh / max | offered (about 12 and 27 cents of output at 1024) | not offered |
Per 1,000 images
- Low: $10.00 against $40.00.
- Medium: $20.00 against $80.00.
- Default settings: about $70.00 (GPT, high) against $80.00 (Ideogram, medium).
- High: $60.00 to $70.00 against $280.00.
Features that are not in the price
ChatGPT Image 2.5 takes up to 16 reference images, an optional mask_url, and background: auto|transparent|opaque. Ideogram 4.5 edits the first reference image and uses up to 4 more, so 5 in total, and selects resolution: 1K|2K. If you need a transparent background or a mask, the first list is the one that documents it. Neither page tells you which model draws the better image for your subject.
A decision rule
Draft on ChatGPT Image 2.5 at low, because exploration at 1 cent is cheap enough to do 40 times for the price of one Ideogram high image. Move to Ideogram 4.5 when you want to compare its output for a specific brief, and pay for it knowingly. In either case set quality in every request on the Image API, so a missing field never decides the bill.
Here is the arithmetic for a mixed batch. Suppose you want 40 drafts and 5 finals. On ChatGPT Image 2.5 at low for drafts and high for finals, that is 40 x 1 cent plus 5 x 7 cents, or 75 cents. On Ideogram 4.5 at low and high, it is 40 x 4 cents plus 5 x 28 cents, or $3.00. The same batch with both set to their defaults costs 45 x 7 cents, $3.15, on GPT and 45 x 8 cents, $3.60, on Ideogram. The gap between a deliberate low and an accidental default is larger than the gap between the two models, which is why the field matters more than the brand.
Rounding also matters. Each image is rounded up to the cent after the 1.25 multiplier, so an image listed at $0.03 bills 4 cents, not 3.75. The catalog figures above are the ones Sume documents; the arithmetic is mine.
Sources
Related posts
More in Comparisons
- ChatGPT Image 2 or 2.5 at high quality: 27 cents vs 7 on Sume
Sume lists ChatGPT Image 2 at a high 1024 rate of $0.211 (27 cents billable) and Image 2.5 at $0.0527 (7 cents). The docs describe xhigh and max for 2.5.
- Cheapest plan with a commercial license: Pika, ElevenLabs, Suno, Luma
ElevenLabs Starter is $6, Suno Pro $8, Luma Plus $30 and Pika Creator $35 on their pages. Free tiers lack commercial use. Sume bills per job.
- Claude MCP connector or Sume Agent Completions for a long video job?
A Messages API request with the MCP connector holds open while Claude works. Sume Agent Completions return 202 and a run to poll. Which one fits a long job.
- Cost per viewer of a 20 second avatar clip at 10, 100 and 1,000 views
A rendered Sume avatar clip costs the same however many watch: $3.68 on standard. That is $0.368 per viewer at 10, $0.0368 at 100, $0.00368 at 1,000.
Written by Sume