Directory
Image models
Text-to-image and image-editing models, with arena Elo where Artificial Analysis or LMArena rank them. 14 image models from 7 developers, 14 marked Notable.
Nano Banana 2.1
Google's newest Flash-tier Gemini image model for generation and conversational editing, released as the replacement for Nano Banana 2.
- Released
- Oct 6, 2026
- $ / image
- $0.034
- Max res
- 4096x4096
FLUX 3 Image
The image mode of Black Forest Labs' FLUX 3 multimodal model: text-to-image, targeted editing and multi-reference composition through one API endpoint.
- Released
- Oct 1, 2026
- $ / image
- $0.048
- Max res
- —
GPT Image 2.5 Flare
OpenAI's default image model from the Images 2.5 launch, the faster sibling of Sunburst for everyday high-quality generation and editing.
- Released
- Sep 8, 2026
- $ / image
- —
- Max res
- 3840x2160
GPT Image 2.5 Sunburst
OpenAI's precision-tier image model from the Images 2.5 launch, aimed at detailed creative work and tight control across edits, at the cost of longer generation times.
- Released
- Sep 8, 2026
- $ / image
- —
- Max res
- 3840x2160
MAI-Image-2.6
Microsoft AI's in-house flagship image model for generation and image-to-image editing, sold through Microsoft Foundry.
- Released
- Aug 10, 2026
- $ / image
- —
- Max res
- 1536x1536
Grok Imagine Image 2.0
xAI's second-generation image model, built around editing, that powers Quality Mode in Grok Imagine and is available in the API.
- Released
- Aug 7, 2026
- $ / image
- $0.040
- Max res
- —
Reve 2.1
Reve's layout-first image model: it plans a structured scene layout before rendering, so individual elements can be edited and re-rendered.
- Released
- Jul 9, 2026
- $ / image
- —
- Max res
- 4096x4096
Muse Image
The first image generation model from Meta Superintelligence Labs. It powers image creation in Meta AI, Instagram and WhatsApp and pairs with Muse Spark to plan, research and compose before rendering.
- Released
- Jul 7, 2026
- $ / image
- $0.010
- Max res
- —
Nano Banana 2 Lite
Google's lowest-cost, lowest-latency Gemini image model (Gemini 3.1 Flash-Lite Image) for high-volume generation and editing.
- Released
- Jun 30, 2026
- $ / image
- $0.034
- Max res
- 1024x1024
Reve 2.0
Reve's first layout-first image model: each image is built from a structured map of objects, text and regions that can be edited individually. Superseded by Reve 2.1.
- Released
- Jun 3, 2026
- $ / image
- —
- Max res
- —
GPT Image 2
The OpenAI image model behind ChatGPT Images 2.0, for generation and editing with flexible sizes and high-fidelity image inputs. Superseded by GPT Image 2.5.
- Released
- Apr 21, 2026
- $ / image
- $0.21
- Max res
- 3840x2160
Nano Banana 2
Google's Flash-tier Gemini image model (Gemini 3.1 Flash Image) for fast, high-volume generation and editing. Deprecated in favour of Nano Banana 2.1.
- Released
- Feb 26, 2026
- $ / image
- $0.067
- Max res
- 4096x4096
GPT Image 1.5
OpenAI's previous-generation image model, behind the December 2025 ChatGPT Images relaunch, with more precise edits and better instruction following.
- Released
- Dec 16, 2025
- $ / image
- $0.13
- Max res
- 1536x1024
Nano Banana Pro
Google's premium Gemini image model (Gemini 3 Pro Image), built on Gemini 3 for studio-quality, precisely controlled generation and editing.
- Released
- Nov 20, 2025
- $ / image
- $0.13
- Max res
- 4096x4096