flux-2/pro/text-to-image
Create reliable, studio-grade visuals with precise color and layout control.
Flux 2 Klein 9B is a compact, step‑distilled text-to-image generator that accepts a prompt and optionally reference images for unified generation and editing. It outputs high-quality still images with strong prompt adherence and low latency, well-suited for rapid creative iteration, product visuals, and interactive design workflows.
Output format: resolution (commonly 1024×1024) / fps (N/A) / duration (N/A) / aspect ratio (square or custom) / audio (N/A)
| Parameter | Required | Type | Default | Range / Options | Description |
|---|---|---|---|---|---|
| multipart* | Yes (*) | object | — | — | Top-level multipart payload container. |
| body | No | object | — | — | Form-data body holding fields such as prompt and generation settings. |
| contentType | No | string | multipart/form-data | — | Content type for the multipart request.
Common multipart body fields (inside multipart.body):
| Field | Required | Type | Default | Range / Options | Description |
|---|---|---|---|---|---|
| prompt* | Yes (*) | string | — | — | Text prompt describing the desired image. |
| steps | No | integer | 25 | Typically 4–50 | Number of inference steps; Flux 2 Klein 9B is distilled to work well at very low steps. |
| width | No | integer | 1024 | Commonly 512–1536 | Output width in pixels. |
| height | No | integer | 1024 | Commonly 512–1536 | Output height in pixels.
| Billing unit | Price |
|---|---|
| Per generated image | $0.003 per image. |
1) Select the model: In RunComfy, choose Flux 2 Klein 9B Text to Image from the Models directory.
2) Write your prompt: Describe subject, style, lighting, camera, and composition; keep it concise but specific for best results with Flux 2 Klein 9B.
3) Optional references: Upload one or more reference images to guide editing or style transfer; match aspect ratios when possible.
4) Set resolution: Start at 1024×1024 for speed; adjust width/height to match your target aspect ratio.
5) Choose steps: Begin with 4–8 steps for fast iteration; increase toward ~25 steps only if you need extra refinement.
6) Run and review: Generate, inspect details, and refine the prompt or references; Flux 2 Klein 9B responds well to incremental prompt edits.
7) Reproduce results: If a seed control is available in your RunComfy session, lock a seed to iterate consistently across variations.
8) API usage: Use the API panel on this page to copy a curl snippet; provide prompt, steps, width, and height as multipart form fields—no self‑hosting required.
Create reliable, studio-grade visuals with precise color and layout control.
Transform visuals with smart region edits and multi-image blending for precise, high-fidelity results.
Create consistent visual stories with advanced image editing and multi-scene control.
Generate images from text prompts with Wan 2.5 Preview.
Change an image’s aspect ratio cleanly with Ideogram 3 Reframe.
Advanced image-to-image tool with geometry-aware edits and consistent identity control for creative workflows.
Flux 2 Klein 9B text-to-image currently supports up to around 2K output resolution (roughly 2–4 megapixels) with a maximum aspect ratio of 16:9. The system is optimized for low-latency inference, rather than ultra-high-resolution output. Prompt length generally should not exceed 512 tokens for stable results.
Flux 2 Klein 9B text-to-image supports up to four reference images when used through the RunComfy Playground or API, leveraging its multi-reference editing workflow. Adding more references may increase latency or reduce fidelity depending on GPU resources.
After testing successfully in the RunComfy Playground, you can transition Flux 2 Klein 9B text-to-image pipelines into production using the RunComfy API. Simply obtain an API key from your workspace dashboard, mirror your prompt and reference parameters, and consume usd tokens per generation request. Refer to RunComfy’s developer documentation for endpoint schemas and rate limits.
Flux 2 Klein 9B text-to-image excels in achieving high-quality results in under a second while maintaining strong prompt adherence. Compared to earlier Flux versions, it runs with only four inference steps, consuming less VRAM but delivering sharper details and more balanced tones.
While Flux 2 Klein 9B text-to-image produces lower maximum resolutions than Seedream 4.5 or Nano Banana Pro (usually capped around 2K), it significantly outperforms them in latency and responsiveness. Its speed makes it ideal for interactive workflows, previsualization, and live art generation, where near-instant feedback is critical.
Yes. Flux 2 Klein 9B text-to-image employs an 8B Qwen3 text embedder, which noticeably improves the fidelity of rendered text and fine details within generated images. This enhancement reduces common distortions in typography seen in earlier models.
Flux 2 Klein 9B text-to-image was designed for ultra-low latency, with sub-second generation possible on advanced GPUs. Its step-distilled architecture (four-step inference) minimizes delay, making it ideal for live previews, interactive sessions, and rapid iteration design use cases.
RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.
