Generation
Available Models
Tesni offers 45+ AI models across multiple providers. All models route through fal.ai using your own API key - you only pay for what you generate.
Text-to-Image Models
Generate images from text prompts. Use with Scene and Image nodes.
| Model | Provider | Tier | ~Cost | Best For |
|---|---|---|---|---|
| Recraft V4 ★ | Recraft | Quality | $0.04 | UI design, typography, layout composition, colour fidelity |
| Recraft V3 | Recraft | Balanced | $0.04 | Vector illustrations, brand palette support |
| FLUX.2 [flex] | Black Forest Labs | Balanced | $0.05 | Best text rendering - headlines, buttons, labels |
| FLUX Kontext [pro] | Black Forest Labs | Quality | $0.04 | State-of-the-art prompt following, flawless typography |
| FLUX Schnell | Black Forest Labs | Fast | $0.003 | Ultra-fast previews (~2s). Rapid iteration |
| Ideogram V3 | Ideogram | Balanced | $0.04 | Exceptional text-in-image rendering, logos |
| Nano Banana 2 | Quality | $0.04 | Gemini 3.1 Flash Image. 4K, web search, 4x faster than Pro | |
| Seedream V4 | ByteDance | Balanced | $0.03 | Strong composition, cinematic quality, low cost |
| Seedream 5.0 Lite | ByteDance | Balanced | $0.03 | Up to 3K resolution, strong prompt adherence |
Image-to-Image (Edit / Remix) Models
Edit existing images using natural language. Use with connected Image nodes.
| Model | Provider | Tier | ~Cost | Best For |
|---|---|---|---|---|
| FLUX Kontext [pro] Edit ★ | Black Forest Labs | Quality | $0.04 | Best contextual editor. Preserves layout while changing elements |
| FLUX Kontext [max] Edit | Black Forest Labs | Quality | $0.08 | Premium editing with improved prompt adherence |
| Nano Banana 2 Edit | Quality | $0.04 | Natural language edits without masks. 4K support | |
| Gemini 2.0 Flash | Fast | $0.01 | Ultra-cheap, lightning-fast iterative edits | |
| Seedream V5 Lite Edit | ByteDance | Balanced | $0.035 | Multi-reference editing, combining design elements |
UGC Character Models
Maintain consistent character identity across different scenes, poses, and contexts. These models preserve facial features, body proportions, and style when generating new content from a Character node.
| Model | Provider | Tier | ~Cost | Best For |
|---|---|---|---|---|
| FLUX Kontext Character ★ | Black Forest Labs | Quality | $0.04 | Best character consistency. Preserves identity across scenes |
| FLUX Kontext [max] Character | Black Forest Labs | Quality | $0.08 | Premium character consistency with better detail |
| FLUX PuLID | Black Forest Labs | Balanced | $0.03 | Face preservation for portraits and headshots |
| FLUX Redux | Black Forest Labs | Balanced | $0.05 | Style transfer while maintaining composition |
| FLUX Kontext Multi [pro] ✨ | Black Forest Labs | Quality | $0.04 | Multi-image composition - combine character + product in one call |
| FLUX Kontext Multi [max] | Black Forest Labs | Quality | $0.08 | Premium multi-image for complex character + product + scene |
💡 Multi-Image Models — Image Order Matters
When using any multi-image model (Kontext Multi, Nano Banana 2 Edit, or Seedream V5 Lite Edit), images are mapped sequentially: the first connected image = "image 1", the second = "image 2". Always put your primary subject (character / base image) as image 1 and the clothing, product, or design reference as image 2. Swapping the order will produce poor results. See the Prompt Writing Guide for detailed tips.
Video Generation Models
Generate videos from images. Connect a source Image node to a Video node and describe the motion you want. Video costs are typically per-second of output.
| Model | Provider | Tier | ~Cost | Best For |
|---|---|---|---|---|
| Kling 1.5 Pro ★ | Kuaishou | Quality | ~$0.10/5s | Reliable high-quality video. 5-10s clips with natural motion |
| Kling 1.6 Pro | Kuaishou | Quality | ~$0.12/5s | Improved motion and longer clip support |
| Kling 2.0 Master | Kuaishou | Quality | ~$0.15/5s | Highest fidelity Kling motion and detail |
| Kling 3.0 v3 Pro ✨ | Kuaishou | Quality | ~$0.14/5s | Character-aware video with @Element tags for product refs |
| Kling O3 ✨ | Kuaishou | Quality | ~$0.16/5s | Newest flagship. 3-15s, native audio, multi-shot |
| Kling 2.5 Turbo | Kuaishou | Fast | ~$0.07/s | Fast iteration. Quick previews at lowest Kling price |
| MiniMax Video | MiniMax | Balanced | ~$0.08/5s | Fast video with good motion quality (max 6s) |
| Luma Ray 2 | Luma AI | Balanced | ~$0.06/5s | Smooth camera motion, cinematic quality |
| Veo 3.1 | Quality | ~$0.12/5s | Google's latest video model, cinematic output | |
| Veo 3 ✨ | Quality | ~$0.40/s | Premium with native audio generation. Top-tier cinematic | |
| Wan 2.1 I2V | Alibaba | Balanced | ~$0.06/5s | Strong motion consistency, good value |
| Wan 2.5 I2V | Alibaba | Balanced | ~$0.05/s | Improved motion over 2.1, great value |
| Hunyuan Video | Tencent | Balanced | ~$0.08/5s | Strong motion and consistency |
| PixVerse V5 | PixVerse | Balanced | ~$0.08/5s | Cinematic motion, good alternative to Kling |
| Sora 2 ✨ | OpenAI | Quality | ~$0.20/s | Exceptional motion, physics, and detail. Premium pricing |
Motion Transfer Models
Re-render your AI character with motion from a reference video. Connect a Character image and a driving video to transfer pose, movement, and expressions.
| Model | Provider | Tier | ~Cost | Best For |
|---|---|---|---|---|
| Kling 3.0 v3 Pro Motion ✨ | Kuaishou | Quality | ~$0.168/s | Best motion transfer quality. Up to 30s. Supports @Element facial consistency |
| Kling 3.0 v3 Std Motion ✨ | Kuaishou | Balanced | ~$0.126/s | Fast motion transfer with v3 quality. Up to 30s |
| Kling 2.6 Pro Motion | Kuaishou | Quality | ~$0.112/s | Character re-render with reference video motion. Up to 10s |
| Kling 2.6 Std Motion | Kuaishou | Fast | ~$0.07/s | Fast character re-render with reference video motion. Up to 10s |
| DreamActor V2 | ByteDance | Balanced | ~$0.05/s | Multi-character, non-human, lip/expression transfer. Up to 30s |
Utility Models
Post-processing tools for upscaling, face recovery, background removal, and inpainting. Select these from the model dropdown on any Image node.
| Model | Provider | ~Cost | Purpose |
|---|---|---|---|
| Bria RMBG 2.0 | Bria | $0.001 | Background removal — clean cutouts for product shots. No prompt needed |
| Face Enhance (Topaz) | Topaz | $0.04 | 2x upscale with AI face recovery. Sharpens facial features in portraits. No prompt needed |
| Clarity Upscaler | fal.ai | $0.02 | AI upscaling with detail enhancement (2x). Accepts an optional prompt to guide enhancement |
| CCSR Upscaler | fal.ai | $0.02 | Content-consistent super-resolution (2x). Best for restoring fine detail. No prompt needed |
| Flux Fill Inpaint | Black Forest Labs | $0.04 | Fill or replace regions of an image with context-aware content. Describe what to fill in the prompt |
💡 How to Use Utility Models
- Open the model picker on any Image node and scroll to the Utility section.
- Connect an upstream image — utility models need at least one source image to process.
- Face Enhance, CCSR Upscaler, and RMBG ignore the prompt — just connect your image and hit Generate.
- Clarity Upscaler accepts an optional prompt to guide the enhancement (e.g. "sharp portrait, detailed skin texture").
- Flux Fill Inpaint uses the prompt to decide what to paint in the masked/empty region.
Pricing Notes
- All costs are approximate per-generation via fal.ai. Actual costs depend on resolution, duration, and model load.
- Image models: Cost is per image generated.
- Video models: Cost is typically per second of output video.
- Utility models: Most are under $0.05 per use.
- ★ = Default / Recommended model. ✨ = Newly added model.
- You bring your own fal.ai API key - Tesni does not mark up or add surcharges.
Learn More
- Pipelines - How to combine models for Character + Product workflows
- Prompt Writing - Tips for better results
- Platforms - Output dimensions for each social platform