Nano Banana 2 by Google — fast image generation across 4 resolution tiers (0.5K–4K). Optional web search and thinking modes. SynthID watermarking on every output.
Two variants for different quality and cost needs — all on the same Runflow API.
Standard — fast generation from 0.5K to 4K resolution
Premium — higher quality, web search enabled, up to 4K
Every model on Runflow is scored per vertical — not generic leaderboard numbers. Scores come from domain-expert evaluation on real production prompts.
| Use Case | Assessment | Notes |
|---|---|---|
| Portrait | Strong | Fast generation with natural skin tones and detail |
| Fashion | Moderate | Good quality across resolution tiers |
| Product | Strong | Clean rendering with SynthID watermarking |
| Creative | Strong | Flexible with optional thinking mode for complex prompts |
| Quality | Price per image | Multiplier |
|---|---|---|
| 0.5K (512px) | $0.060 | 0.75× |
| 1K (standard) | $0.080 | 1× |
| 2K | $0.120 | 1.5× |
| 4K | $0.160 | 2× |
$1.00 runs this model ~12 times at standard rate. Pricing is subject to change.
Full parameter reference for the nano-banana-2 model endpoint.
| Parameter | Type | Description |
|---|---|---|
| prompt | string | Text prompt describing the image. 3–50,000 characters. |
| resolution | string | Output resolution. Options: "0.5K", "1K", "2K", "4K". Default: "1K". Price varies by tier. |
| aspect_ratio | string | Image aspect ratio. Options: "21:9", "16:9", "3:2", "4:3", "5:4", "1:1", "4:5", "3:4", "2:3", "9:16", "4:1", "1:4", "8:1", "1:8". Use "auto" to let the model decide. Default: "auto". |
| enable_web_search | boolean | Enable web search to ground generation in current information. +$0.015/image. Default: false. |
| thinking_level | string | Enable model thinking. Options: "minimal", "high". Adds +$0.002/image at high level. Omit to disable. |
| num_images | integer | Number of images to generate. Range: 1–4. Default: 1. |
| output_format | string | Output format. Options: "jpeg", "png", "webp". Default: "png". |
| safety_tolerance | string | Content moderation level. 1 = most strict, 6 = least strict. Default: 4. |
| seed | integer | Random seed for reproducible outputs. |
| limit_generations | boolean | Limit to 1 image per prompt round regardless of prompt instructions. Default: true. |
| sync_mode | boolean | Return media as data URI without storing in history. Default: false. |
Flux AI models are available on many platforms. Here's what makes the Runflow image generation API different.
From 0.5K (fastest, cheapest) to 4K — choose the right resolution per request without switching models.
Native support for 14 ratios from 21:9 to 1:8 — including panoramic (8:1) and ultra-tall (1:8) formats with no letterboxing.
Maintain consistent character appearance across up to 5 people in a single generation — ideal for storytelling, campaigns, and multi-character scenes.
Every output carries Google's SynthID digital watermark, invisibly embedded and robust to common edits — enabling attribution and provenance tracking.
Optional web search grounds generation in current events. Thinking mode adds reasoning for complex prompts — both priced as add-ons.
Every output is scored by Runflow's Sentinel system before delivery, catching low-quality results automatically.
Same API, different strengths. Switch models with one parameter change.
Black Forest Labs
FLUX.1 [dev] is a 12B-parameter text-to-image model by Black Forest Labs. Open-w…
ByteDance
Seedream 4.5 by ByteDance — unified generation and editing in one model. Strong…
ByteDance
Seedream 4 by ByteDance — unified generation and editing model with photorealism…
OpenAI
GPT-Image 1.5 by OpenAI — high-fidelity images with strong prompt adherence. Thr…
Nano Banana Pro by Google — premium image generation with sharp detail and natur…
Alibaba
Qwen Image 2 by Alibaba — 7B-parameter model for photorealism and accurate text…
Get API access in minutes. No GPU setup, no infrastructure to manage. Pair with Sentinel to control the quality of every AI-generated image you ship.