AI image model news
Last updated
New text-to-image, AI photo editing and upscaling models ship almost every week. This page tracks the releases that matter, newest first, each with a link to the original announcement. For the full list of models you can use today, see the live AI image model directory.
Trending this month
- Nano Banana 2.1
- Nano Banana 2
- Ideogram 4.5
- FLUX 3 Image Klein
- Krea 3
- Nano Banana Pro
Ideogram releases Ideogram 4.5 edit model with reduced edit drift
Ideogram added Ideogram 4.5, an edit model that reduces drift across multi-turn edits, preserving details and handling precise color and restoration changes. It supports high-resolution edits without downsizing, preserving crop edges so edits can be stitched back seamlessly. Ideogram also added AI background removal that turns uploads into transparent PNGs in the app and via API.
Black Forest Labs releases FLUX 3 Image Klein, a 5.6B open-weights image model
Black Forest Labs released FLUX 3 Image Klein, a 5.6B-parameter open-weights text-to-image model supporting image editing and multi-reference generation with up to 10 reference images. It features native 4K output, pixel-perfect editing and bounding box controls, and runs on GPUs with as little as 8GB VRAM.
Krea launches Krea 3 beta with 2K image edits in under 8 seconds
Krea opened the beta for Krea 3, an image editing model that delivers end-to-end 2K edits in under 8 seconds. It matches the quality of Nano Banana Pro and GPT Image 2.5 and supports bounding box control with precise multi-turn edits for iterative creative workflows. A full release with additional benchmarks is planned for the following week.
Artificial Analysis: Nano Banana 2.1 ranks #4 on text-to-image and editing benchmarks
Artificial Analysis evaluations published October 9, 2026 found Nano Banana 2.1 replaces Nano Banana 2 and ranks #4 on both AA-Image-T2I v2.0 and AA-Image-Editing v2.0 at half its predecessor's price. It improved on Nano Banana 2 in all 9 text-to-image capabilities and became Google's best model on all 7 editing actions.
Google releases Nano Banana 2.1 image model, deprecates Nano Banana 2
Google released Nano Banana 2.1 on October 6, 2026 as gemini-nano-banana-2.1, built on Gemini 3.6 Flash, and marked Nano Banana 2 as deprecated the same day. It costs $0.0336 per 1K image (half of Nano Banana 2) and Google's own evals place it ahead of Nano Banana Pro on every row. It supports up to 14 reference images, 1K/2K/4K output, and text, image, video and PDF inputs.
Alibaba's Qwen Image 3.0 and Qwen Image 3.0 Pro available via EvoLink API
Qwen Image 3.0 and Qwen Image 3.0 Pro are available through EvoLink's unified image API. The release supports text-to-image, 1-3 reference-image editing, text-rich layouts, multilingual text, batch output of 1-6 images, up to 4.5K-token prompts, 10px text and native rendering in 12 languages.
Ideogram 4.5 adds precision multi-turn image editing
Ideogram released Ideogram 4.5, a precision editing model for multi-turn refinement and reference-guided edits, designed to make high-resolution in-place changes with fewer pixel, color and texture shifts.
Qwen-Image-2.1 released as an open image model
Qwen-Image-2.1 was released on September 20, 2026 and is listed as an open-weights image generation model.
OpenAI releases GPT Image 2.5 (ChatGPT Images 2.5)
OpenAI released GPT Image 2.5, known as ChatGPT Images 2.5, on September 8, 2026. It adds a Sketch feature that turns drawings into AI images and cuts image generation latency by about 50%.
LLaDA-Image and LLaDA-Image Turbo released as open image models
LLaDA-Image and a faster LLaDA-Image Turbo variant were released on September 4, 2026 and are listed as open image generation models.
MAI-Image-2.6-Flash joins Microsoft's MAI-Image line
MAI-Image-2.6-Flash was released on September 4, 2026, a faster variant in the MAI-Image-2.6 family, which launched in August 2026.
Try the top models in PixelMagic
Generate, edit and upscale with leading AI image models from one account and one credit balance.
Start creating free