Chroma
Open-weights text-to-image model at $0.01 per image with zero-retention privacy.
Generate imageGet API key- Provider
- Lodestones
- Price
- $0.01 / image
- Prompt limit
- 7,500 chars
- Released
- Jan 2026
- License
- Open weights
What is Chroma?
Chroma is an open-weights text-to-image model distributed via Hugging Face by Lodestones. Available on Venice since January 2026, it generates images from text prompts up to 7,500 characters. As an open-source image model with flat per-image pricing, it offers a private, zero-retention alternative to closed Big-Tech image generators.
Use it privately on Venice
On Venice, Chroma runs under a private, zero-retention privacy tier — your prompts and generations are not stored or used for training. Because Venice hosts the open weights directly, you get permissionless access to an open-source image model without surrendering your creative data to centralized surveillance platforms. It is also featured in Venice's Recommended and Featured collections.
What can it do?
- •Open weights — you can inspect, download, or self-host the model independently of Venice.
- •Extremely cost-effective flat pricing at $0.01 per image with no subscription lock-in.
- •Private by design on Venice — prompts are processed with zero retention and not used for training.
- •Generous 7,500-character prompt limit for detailed scene composition.
- •Recognized by Venice in its Recommended and Featured collections.
- •Not uncensored — it will decline certain prompts according to its safety configuration.
- •No advanced capabilities such as tool use, web search, vision input, or reasoning are available.
- •As a community open-weights model, it may lag behind closed frontier models in photorealistic detail and prompt adherence.
- •Venice does not run this model in a TEE or with end-to-end encryption; privacy relies on policy-grade zero retention.
Sample outputs
Generated on Venice with our standard prompt suite — the same prompts we run through every model of this type, so you can judge it like-for-like.
A retro travel poster with the bold headline "VENICE" in large condensed serif type, sunset color palette, clean layout
Photorealistic close-up portrait of a weathered fisherman at golden hour, 85mm lens, shallow depth of field, natural skin texture
A small red cube balanced on top of a large glossy blue sphere, with a green cone to the right, plain light-grey studio background
Cozy watercolor illustration of a hillside village in autumn, warm tones, soft paper texture
How to use it via API
Venice exposes this model through the REST API. Queue a generation with chroma.
curl https://api.venice.ai/api/v1/image/generate \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "chroma",
"prompt": "A serene mountain lake at dawn, photorealistic"
}' --output image.pngSpecifications
Pricing
Flat per-image pricing on Venice: $0.01 per generation.
New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.
Chroma vs alternatives
| Model | Price (Venice) | Open weights | Prompt limit | Privacy on Venice |
|---|---|---|---|---|
| Chroma | $0.01 / image | Yes | 7,500 chars | Zero retention |
| Lustify v7 | $0.01 / image | No | — | — |
| Flux 2 Max | $0.09 / image | No | — | — |
| Grok Imagine | from $0.04 / image | No | — | — |
The only open-weights option in this group, priced at a flat $0.01 per image with zero prompt retention.
What is it good for?
- •General text-to-image generation for creative prototyping and exploration.
- •Bulk image generation workflows where low per-image cost matters.
- •Users and teams who prefer an open-source stack and model sovereignty.
- •Privacy-conscious creators who want to avoid centralized data collection.
Prompting tips
- •Stay within the 7,500-character limit to avoid truncation.
- •Be explicit about subject, style, and lighting — open models often benefit from detailed descriptions.
- •Use the 2× or 4× upscale options ($0.02 and $0.08) if you need higher resolution output.
Frequently asked questions
Chroma is an open-weights text-to-image model distributed by Lodestones via Hugging Face. It generates images from text prompts on Venice for a flat per-image fee.
Chroma costs $0.01 per image. Optional upscaling costs $0.02 for 2× and $0.08 for 4×. There is no subscription required.
Yes. Chroma is an open-weights model with publicly available weights on Hugging Face, meaning you can inspect, download, or self-host it independently of Venice.
On Venice you pay per generation; it is not free beyond any welcome credits included with your account. Because the weights are open, you can also run it locally for free if you have the hardware.
Chroma is open weights and costs $0.01 per image, making it ideal for open-source workflows and bulk generation. Flux 2 Max is a closed model at $0.09 per image that may offer stronger photorealism. Choose Chroma for sovereignty and cost; choose Flux for top-tier closed-model quality.
No. Chroma does not support tool use, vision input, reasoning, or web search. It is a standalone text-to-image generation model.
Venice processes Chroma generations under a private, zero-retention policy. Your prompts and images are not stored, profiled, or used for training. Note that Venice does not currently run image models in a TEE or with end-to-end encryption.
You can use up to 7,500 characters per prompt on Venice.
No. While Chroma is open weights, its safety configuration means it will still refuse certain prompts. It is not marketed as an uncensored model.
Related models
Run Chroma privately.
No prompt logging. No data used for training. Free to start — no credit card.
