SANA Sprint 1.6B
Model Information
Display Name: SANA Sprint 1.6B
API Model ID: nvidia/sana-sprint-1.6b
Category: Text To Image
Description: SANA-Sprint 1.6B is NVIDIA's ultra-efficient text-to-image model that generates 1024px images in just 1-4 inference steps using continuous-time consistency distillation. **Key Features:** - 1.6B parameters with one-step generation capability - 1024px output at ~0.1s per image on H100 - 10x faster than comparable diffusion models - Continuous-time consistency distillation (sCM) - BF16 precision with DC-AE latent encoding (32x compression) **Capabilities:** - Ultra-fast text-to-image generation - High-quality 1024x1024 output in 1-4 steps - Multi-aspect ratio support - Adjustable guidance scale **Best For:** - Real-time image generation applications - High-throughput production pipelines - Low-latency interactive use cases - Cost-effective batch generation **Technical Specs:** - Architecture: Linear diffusion transformer - Parameters: 1.6 billion - Text Encoder: Gemma-2-2B-IT - Latent Encoder: DC-AE (32x spatial compression) - FID Score: 7.59 (single step) - GenEval: 0.74 - License: NSCL v2-custom (NVIDIA)
How to Use This Model
To use SANA Sprint 1.6B via the HInow.ai API, use the model ID: nvidia/sana-sprint-1.6b
API Request Example (Chat/Text)
POST https://api.hinow.ai/v1/chat/completions
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
{
"model": "nvidia/sana-sprint-1.6b",
"messages": [
{"role": "user", "content": "Your message here"}
]
}
API Request Example (Image Generation)
POST https://api.hinow.ai/v1/images
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
{
"model": "nvidia/sana-sprint-1.6b",
"prompt": "Your image description here"
}
Pricing
- total: $0.005
Available Parameters
- aspect_ratio: Output image aspect ratio (Options: 1:1, 4:3, 3:4, 16:9, 9:16)
- num_inference_steps: Number of inference steps (default: 2) (Options: 1, 2, 3, 4)
- guidance_scale: Guidance scale (default: 4.5) (Options: 1.0, 2.0, 3.0, 4.5, 6.0, 8.0)
- output_format: Output image format (Options: jpeg, png)
Quick Reference
To use this model, set: "model": "nvidia/sana-sprint-1.6b"
Featured: No
Documentation: https://hinow.ai/models/nvidia/sana-sprint-1.6b
API Endpoint: https://api.hinow.ai/v1
