- 🆕 Open-source 8.9B text-to-image model built on FLUX.1-schnell.
- 🔒 Fully Apache 2.0 licensed for unrestricted commercial use.
- 🎯 Tuned for vivid color saturation and controlled palettes.
- 📚 Trained on diverse data including photography, anime and furry art.
- 🌐 Community-led development by Lodestones on Hugging Face.
- 👁️ Available in Base and HD variants.
- 🧠 Rectified-flow transformer architecture derived from FLUX.
- ⚡ Runs in ComfyUI and uses a T5-XXL text encoder.
Community represents the broader ecosystem of independent creators, fine-tuners, and open-source contributors who build and share models outside any single corporate lab. Rather than a formal organization, this category collects specialized models developed by individual…
Explore 4 more models by Community →Chroma is a community-developed, open-source text-to-image model with roughly 8.9 billion parameters, built on the FLUX.1-schnell rectified-flow transformer architecture and released under a permissive Apache 2.0 license. Within the Venice catalog it is positioned as an image generator tuned for vivid color saturation and controlled palettes, complementing other community image models such as Lustify v7, Lustify v8, and Anime (WAI).
Rather than a single checkpoint, Chroma is a family of related releases from developer Lodestones, published on Hugging Face: a core Base model and a 1024x1024 high-resolution HD fine-tune. The project is maintained by the Lodestones organization rather than a single company.
Because Chroma is a foundational base rather than an iteration on an earlier Chroma generation, there is no same-family predecessor to compare against here. Its distinguishing traits, per the project's own materials, are full transparency around its training and a training set spanning anime, furry art, and photography, atop an architecture derived from FLUX.1-schnell.
Practically, Chroma runs in ComfyUI workflows and requires the T5-XXL text encoder and a FLUX VAE alongside the diffusion checkpoint. The Apache 2.0 terms allow free use, modification, redistribution, and commercial deployment without royalties, which makes it suitable for trainers building specialized derivatives.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | $ / img | Categories | API |
|---|---|---|---|---|---|
| Venice.ai Proxy 0x1f22…18c9 | 100.00 | #2 | $0.005 | image,creative | openai-images |
| ▲ Apex Ant 0x73b4…e736 | 91.01 | #1 | $0.004 | image,media | openai-images |
| D5V1N2 0xd5e7…7be0 | 33.17 | gated | $0.012 | image,creative,router,fallback | openai-images |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = on-chain trust (0-100). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.