ERNIE-Image
ERNIE-Image is an open text-to-image generation model developed by Baidu, designed to deliver high-quality visuals with strong instruction accuracy and controllability. It is built on a single-stream Diffusion Transformer (DiT) architecture with around 8 billion parameters, allowing it to achieve state-of-the-art performance among open-weight image models while remaining relatively efficient. The model includes a built-in prompt enhancement system that expands simple user inputs into richer, structured descriptions, improving the quality and consistency of generated images. ERNIE-Image is optimized for complex instruction following, enabling accurate rendering of text within images, structured layouts, and multi-element compositions, making it particularly suitable for use cases like posters, comics, and multi-panel designs. It supports multilingual prompts, including English, Chinese, and Japanese, broadening accessibility and usability across regions.
Learn more
FLUX.1 Krea
FLUX.1 Krea is an open source, guidance-distilled 12 billion-parameter diffusion transformer released by Krea in collaboration with Black Forest Labs, engineered to deliver superior aesthetic control and photorealism while eschewing the generic “AI look.” Fully compatible with the FLUX.1-dev ecosystem, it starts from a raw, untainted base model (flux-dev-raw) rich in world knowledge and employs a two-phase post-training pipeline, supervised fine-tuning on a hand-curated mix of high-quality and synthetic samples, followed by reinforcement learning from human feedback using opinionated preference data, to bias outputs toward a distinct style. By leveraging negative prompts during pre-training, custom loss functions for classifier-free guidance, and targeted preference labels, it achieves significant quality improvements with under one million examples, all without extensive prompting or additional LoRA modules.
Learn more
Gemini 3 Pro Image
Gemini Image Pro is a high-capability, multimodal image-generation and editing system that enables users to create, transform, and refine visuals through natural-language prompts or by combining multiple input images, with support for consistent character and object appearance across edits, precise local transformations (such as background blur, object removal, style transfers or pose changes), and native world-knowledge understanding to ensure context-aware outcomes. It supports multi-image fusion, merging several photo inputs into a cohesive new image, and emphasizes design workflow features such as template-based outputs, brand-asset consistency, and repeated character/person-style appearances across scenes. It includes digital watermarking to tag AI-generated imagery and is available through the Gemini API, Google AI Studio, and Gemini Enterprise Agent Platform.
Learn more
Gemini 3.1 Flash Image
Gemini 3.1 Flash Image is Google DeepMind’s latest image generation model, combining advanced Pro-level capabilities with lightning-fast performance. It delivers enhanced world knowledge, enabling more accurate subject rendering and data-informed visuals grounded in real-time information. The model improves precision text rendering and in-image translation, making it well-suited for marketing assets, infographics, and localized creative content. Stronger instruction following ensures complex prompts are executed with clarity and accuracy. Gemini 3.1 Flash Image maintains subject consistency across multiple characters and objects within a single workflow. It supports production-ready outputs with customizable aspect ratios and resolutions up to 4K. Available across Gemini, Search, AI Studio, Google Cloud, and more, it brings high-quality visual generation at Flash-level speed.
Learn more
