Models¶

Pass --model an alias below or the provider's own model id.
genimg models shows which ones your credentials list. It caches for 5 days; --refresh probes again.

Gemini¶

Alias	Model id	Cost per image by `--resolution`	Aspect ratios	`--thinking`
`gdm:nbp`	`gemini-3-pro-image`	1K $0.134 · 2K $0.134 · 4K $0.24	classic	❌
`gdm:nb2`	`gemini-3.1-flash-image`	512 $0.045 · 1K $0.067 · 2K $0.101 · 4K $0.151	extended	`minimal` `high`
`gdm:nb2-lite`	`gemini-3.1-flash-lite-image`	1K $0.034	extended	`minimal` `high`

Classic aspect ratios: 1:1 2:3 3:2 3:4 4:3 4:5 5:4 9:16 16:9 21:9. Extended adds 1:4 4:1 1:8 8:1.
Every Gemini model accepts --mode batch. Aspect ratio does not change the price.

OpenAI¶

Alias	Model id	Cost per 1K square image by `--quality`
`oai:gi2.5`	`gpt-image-2.5-sunburst`	low $0.006 · medium $0.013 · high $0.053 · xhigh $0.094 · max $0.211
`oai:gi2.5-flare`	`gpt-image-2.5-flare`	low $0.006 · medium $0.013 · high $0.053 · xhigh $0.094 · max $0.211
`oai:gi2`	`gpt-image-2`	low $0.006 · medium $0.053 · high $0.211
`oai:gi1.5`	`gpt-image-1.5`	low $0.009 · medium $0.034 · high $0.133
`oai:gi1`	`gpt-image-1`	low $0.011 · medium $0.042 · high $0.167
`oai:gi1-mini`	`gpt-image-1-mini`	low $0.005 · medium $0.011 · high $0.036

--quality defaults to medium; auto is estimated as medium.

GPT Image 2 and 2.5 take any size in this table; GPT Image 1, 1.5 and 1 mini take only 1K 1:1. genimg rejects any other pair before the request:

`--resolution`	`1:1`	`4:3` / `3:4`	`16:9` / `9:16`
1K	1024×1024	1024×768	❌
2K	2048×2048	2048×1536	2048×1152
4K	2880×2880	❌	3840×2160

On GPT Image 2 and 2.5 a wide 2K image can cost less than a 1K square.
--dry-run prints the estimate for any combination.

Cost follows pixel count and quality: GPT Image 2 and 2.5 count output tokens the way OpenAI's calculator does, and the older models use OpenAI's per-image price for 1024×1024.

$ genimg "a lighthouse at dusk" \
    --model oai:gi2 \
    --resolution 2K \
    --aspect-ratio 16:9 \
    --dry-run
genimg openai/direct oai:gpt-image-2 → gpt-image-2
  prompt   "a lighthouse at dusk"
  params   n=1 q=medium (default) r=2K a=16:9 → 2048x1152
  cost     $0.0424 (estimate)  id=20260925_114713_d61db9
  output   ~/.genimg/generations/20260925_114713_d61db9.png
dry-run: no API call made.

Estimates exclude text input, under a cent per prompt, and apply to Azure too. genimg models --refresh shows which ids your deployment serves.

Codex¶

Alias	What runs	Controls	Cost
`codex:image`	Codex image generation; Codex picks the model and size	`--aspect-ratio` only, as a request in the prompt	Your ChatGPT plan's Codex allowance

genimg records an API-equivalent price range for Codex runs, for comparison only. See Codex subscription.

Other ids¶

An id missing from the tables still works if it matches gpt-image-* (OpenAI) or gemini-*-image* (Google), dated snapshots included.
gdm:imagen4 errors and points to gdm:nb2, because Google retired Imagen 4.

Prices last checked¶

Provider	Source	Checked
Google	Gemini API pricing, Vertex AI pricing	25 September 2026
OpenAI	API pricing, image cost calculator	24 September 2026
Codex	Codex pricing	25 September 2026

Prices are estimates. Your bill depends on tier, batch discounts and the exact pixels returned.