Short answer
- Need animated GIFs or SVG files? GPT Image 2.6 — it's the only one of these whose official docs list GIF and SVG output.
- Lowest list price per image? Seedream 5.0 flash ($0.018), then Qwen Image 3.0 ($0.03) and Nano Banana 2.1 (about $0.034 at 1K).
- Dense text, posters and complex instructions? Nano Banana Pro (Google's premium model with reasoning) or Qwen Image 3.0 (Alibaba cites 10px text in 12 languages).
- Multi-step edits that keep details? GPT Image 2.6 Sunburst or Nano Banana 2.1 — both vendors call out consistency across edits.
- One key for several of them? On Moyi API you can call GPT Image 2.6 (Flare and Sunburst) and Google's Nano Banana 2.1, Nano Banana Pro and Nano Banana 2 with the same key. Seedream and Qwen Image aren't on Moyi API; you'd use BytePlus or Alibaba Cloud directly.
Everything below comes from each vendor's own documentation and pricing pages, checked on November 5, 2026.
New accounts start with a $0 balance — top up before your first request. Top up from $5. If a request returns 402 (400 on the OpenAI format) with "wallet balance is insufficient", your balance is empty — your base URL and key are fine. Top up and retry.
Get one API key for GPT Image 2.6 and Nano Banana — sign up and top up from $5 by card.
The models at a glance
| Model | Vendor | API model ID | What the vendor says it's for | On Moyi API |
|---|---|---|---|---|
| GPT Image 2.6 Flare | OpenAI | gpt-image-2.6-flare |
Fast iteration; as little as ~6 s per image at low quality | Yes |
| GPT Image 2.6 Sunburst | OpenAI | gpt-image-2.6-sunburst |
Detailed work; more consistent details from edit to edit at medium/high quality | Yes |
| Nano Banana 2.1 | gemini-nano-banana-2.1 |
Google's main high-efficiency model: Flash-level speed, better text rendering and multi-turn consistency, 1K/2K/4K | Yes | |
| Nano Banana Pro | gemini-3-pro-image |
Premium model that uses reasoning for complex instructions, brand consistency and high-fidelity text | Yes | |
| Nano Banana 2 | gemini-3.1-flash-image |
Previous high-efficiency model; Google recommends 2.1 for new projects | Yes | |
| Seedream 5.0 pro / flash / lite | ByteDance (BytePlus) | dola-seedream-5-0-pro-260628, dola-seedream-5-0-flash-260915, seedream-5-0-lite-260128 |
Pro: follows complex instructions with less prompt effort; layer decomposition and interactive editing | No |
| Qwen Image 3.0 / 3.0 Pro | Alibaba Cloud | qwen-image-3.0, qwen-image-3.0-pro |
Prompts up to 4.5k tokens, dense layouts, 10px text in 12 languages | No |
Sources: OpenAI image generation guide and @OpenAIDevs; Google Gemini image generation docs; BytePlus Seedream 5.0 and ModelArk pricing; Alibaba Cloud qwen-image-3.0.
Features only some models have
| GPT Image 2.6 | Nano Banana 2.1 / Pro | Seedream 5.0 | Qwen Image 3.0 | |
|---|---|---|---|---|
| Animated GIF output | Yes (2–64 frames) | Not listed | Not listed | Not listed |
| SVG (vector) output | Yes | Not listed | Not listed | Not listed |
| Planning control | complexity: low / medium / high |
Pro uses reasoning ("Thinking") | — | — |
| Targeted edits | Markup overlay on /v1/images/edits |
Conversational, multi-turn editing | Interactive editing (pro / flash); layer decomposition (pro) | Text-and-image input |
| Up to 4K output | Explicit sizes up to 3840 px on the long side (tested on 2.5) | 1K / 2K / 4K | Higher-than-1.5K tier (pro) | 1K / 2K listed |
| Search grounding | — | Google Search grounding | — | — |
"Not listed" means the vendor's docs we checked don't mention it, not that it's impossible.
Official prices
Per output image unless noted; list prices from each vendor's pricing page, without promotions.
| Model | Output price | Notes |
|---|---|---|
| GPT Image 2.6 Flare / Sunburst | $30 per 1M image output tokens | Text input $5 and image input $8 per 1M tokens; cost per image depends on size and quality (OpenAI pricing) |
| Nano Banana 2.1 | ≈ $0.034 (1K) · $0.050 (2K) · $0.113 (4K) | $30 per 1M image output tokens; input $1.50 per 1M (Google pricing) |
| Nano Banana Pro | ≈ $0.134 (1K/2K) · $0.24 (4K) | $120 per 1M image output tokens; input $2.00 per 1M |
| Nano Banana 2 | ≈ $0.067 (1K) · $0.101 (2K) · $0.151 (4K) | $60 per 1M image output tokens; input $0.50 per 1M |
| Seedream 5.0 pro | $0.045 (up to 1.5K) · $0.09 (above) | First input image free, then $0.003 each |
| Seedream 5.0 flash | $0.018 | — |
| Seedream 5.0 lite | $0.035 | — |
| Qwen Image 3.0 | $0.03 (1K or 2K) | International deployment; input $0.003 per image |
| Qwen Image 3.0 Pro | $0.04 (1K) · $0.075 (2K) | International deployment |
The Google per-image figures are Google's own conversions. OpenAI doesn't publish a fixed per-image price for GPT Image 2.6; use the calculator in its image generation guide or the usage field in the response. Moyi API's own prices are on the pricing page.
Which one to pick
For GIFs, stickers and animated mascots: GPT Image 2.6 Flare to iterate on the motion, Sunburst for the final loop. Nothing else in this list lists GIF output.
For logos and icons you'll edit later: GPT Image 2.6 with output_format: "svg". The others return raster images.
For high volume at the lowest unit price: Seedream 5.0 flash or Qwen Image 3.0 if you're fine using BytePlus or Alibaba Cloud directly; Nano Banana 2.1 if you want it on the same key as your other models.
For posters, infographics and UI mockups with lots of text: Nano Banana Pro or Qwen Image 3.0; GPT Image 2.6 Sunburst with complexity: "high" is the OpenAI option.
For product photos edited in several steps: GPT Image 2.6 Sunburst with a markup overlay, or Nano Banana 2.1 in a multi-turn conversation.
Calling them through one key
GPT Image 2.6 uses the OpenAI Images API format:
curl https://api.moyi-api.com/v1/images/generations \
-H "Authorization: Bearer $MOYI_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "gpt-image-2.6-flare", "prompt": "A sticker of a smiling cactus, flat colors", "quality": "low"}'
Nano Banana uses the native Gemini API (generateContent), with the same Moyi API key:
curl https://api.moyi-api.com/v1beta/models/gemini-nano-banana-2.1:generateContent \
-H "Authorization: Bearer $MOYI_API_KEY" \
-H "content-type: application/json" \
-d '{
"contents": [{"role": "user", "parts": [{"text": "A sticker of a smiling cactus, flat colors"}]}],
"generationConfig": {"responseModalities": ["IMAGE"], "imageConfig": {"imageSize": "1K"}}
}'
Both return the image as Base64. The full parameter lists are in GPT Image 2.6 API and Nano Banana 2.1 API.
FAQ
Is GPT Image 2.6 better than Nano Banana?
It depends on the job. GPT Image 2.6 is the one with GIF and SVG output and a complexity control; Nano Banana 2.1 is cheaper per 1K image at list price and adds Google Search grounding; Nano Banana Pro is Google's pick for complex instructions and text.
Which is the cheapest image API?
At list price per image, Seedream 5.0 flash ($0.018), then Qwen Image 3.0 ($0.03) and Nano Banana 2.1 (about $0.034 at 1K).
Can I use Seedream or Qwen Image through Moyi API?
Not at the moment; they aren't on Moyi API's price list. GPT Image 2.6 and Google's Nano Banana models are.
Do I need an OpenAI or Google account?
Not through Moyi API: one Moyi API key works for both. See how to use the GPT Image 2.6 API without OpenAI organization verification.