Other notable image models

A few additional models that are accessible through Krea, Flora, or the broader ecosystem and worth mentioning briefly:

HiDream

An open-source image model released in 2025 that has gained traction in the ComfyUIComfyUIThe dominant node-based interface for running diffusion models locally. Started by 'comfyanonymous' in January 2023, now stewarded by Comfy Org. community for high-quality outputs and good license terms. Based on a DiTDiT (Diffusion Transformer)A diffusion model architecture that uses a Transformer instead of a U-Net. Introduced by Peebles and Xie in 2022. Scales much better than U-Nets and is the basis for almost all current frontier models. architecture, with a focus on photorealistic generation. Less well-known than FLUX but used by some power users as an alternative.

Hunyuan Image (Tencent)

Tencent's open-source image generation model. Less prominent than its video counterpart Hunyuan Video but available as an alternative in the open-source ecosystem. Tencent's general AI strategy mirrors Alibaba's: release models openly to build cloud infrastructure mindshare in China while also competing internationally.

Playground v3

From a startup called Playground, this was an interesting model that emphasized aesthetic quality and graphic design use cases. The company has since pivoted, but the model is still available through some platforms.

Adobe Firefly

Adobe's proprietary image model, integrated into Photoshop, Illustrator, and other Creative Cloud tools. Trained exclusively on Adobe Stock content (and licensed material), which makes it the safest model from a copyright perspective for commercial use. In early 2026, Adobe introduced a Creative Agent that orchestrates entire creative workflows autonomously, selecting and chaining Firefly tools without manual intervention, and the Firefly Video Model now integrates Generative Fill for Video. These additions signal Adobe's move from individual generation features toward agentic creative production, where the AI handles multi-step workflows rather than single generations. Not accessible through Krea or Flora, but used widely by professional designers because of the IP guarantees and the deep integration with the existing Creative Cloud workflow.

With image models covered, we move to video, which is where the field is moving fastest, and where the architectural and economic constraints are the most extreme.