Latest month in the catalog
New AI image models: November 2026
8 AI image models are dated November 2026 in Versely’s catalog, from 3 providers: OpenAI (4), ByteDance (2) and MiniMax (2). The newest is MiniMax H3 Image, released 10 November 2026.
| Model | Built by | Released | Credits per job |
|---|---|---|---|
| MiniMax H3 Image | MiniMax | 10 November 2026 | 3-7 credits |
| MiniMax H3 Image Edit | MiniMax | 10 November 2026 | 7 credits |
| Seedream 5.5 Editvariant | ByteDance | 9 November 2026 | 5 credits |
| Seedream 5.5 | ByteDance | 9 November 2026 | 3 credits |
| GPT Image 2.6 Sunburst Text to Imagevariant | OpenAI | 4 November 2026 | 5-9 credits |
| GPT Image 2.6 Sunburst Editvariant | OpenAI | 4 November 2026 | 7-11 credits |
| GPT Image 2.6 Flare Text to Imagevariant | OpenAI | 4 November 2026 | 5-9 credits |
| GPT Image 2.6 Flare Edit | OpenAI | 4 November 2026 | 7-11 credits |
See November 2026 in the 2026 image timeline, or jump to every month of 2026.
The 2026 image timeline
1–15 January 2026
5 models · 2 providersFlux and PrunaAI shipped 5 image models between 1 January 2026 and 15 January 2026. Of the 5, 4 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 1 credit on Pruna Image Edit. Highest output in it: 4K on Pruna Image Edit.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Pruna Image EditPrunaAI's efficiency-optimised image editor — instruction-based edits at the single-credit tier, compressed for speed. The… | PrunaAI | 1 credit | 4K | 16:9, 3:2, 4:3, 1:1, 3:4, 2:3 |
| Flux 2 FlashUltra-fast image generation with Flux 2 | Flux | 1 credit | 4K | 1:1, 4:3, 3:4, 16:9, 9:16 |
| Flux 2 Flash TurbovariantTurbo speed image generation with Flux 2 | Flux | 2 credits | 4K | 1:1, 4:3, 3:4, 16:9, 9:16 |
| Flux 2 Klein 4B Base EditFlux 2 Klein 4B Base Edit is a compact yet powerful image editing model built on the Flux 2 architecture. It enables high-quality… | Flux | 1 credit | — | 16:9, 4:3, 1:1, 3:4, 9:16 |
| Flux 2 Klein 9B BaseText-to-image generation with FLUX.2 [klein] 9B Base from Black Forest Labs. Enhanced realism, crisper text generation, and… | Flux | 1 credit | — | 1:1, 4:3, 3:4, 16:9, 9:16 |
26 January – 1 February 2026
3 models · 2 providersHunyuan and PrunaAI shipped 3 image models between 26 January 2026 and 1 February 2026. Of the 3, 2 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on FireRed Image Edit.
- New name on the roster: Hunyuan.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| HunyuanImage 3.0 Instruct EditHunyuanImage 3.0 Instruct Edit is an advanced instruction-following image editing model by Tencent. It allows users to modify and… | Hunyuan | 8 credits | — | 16:9, 4:3, 1:1, 3:4, 9:16, auto |
| HunyuanImage 3.0 Instruct Text to ImagevariantHunyuanImage 3.0 Instruct Text to Image is a powerful text-to-image model by Tencent that generates high-quality images from… | Hunyuan | 8 credits | — | 16:9, 4:3, 1:1, 3:4, 9:16, auto |
| FireRed Image EditAdvanced image editing model for prompt-guided modifications, inpainting, and creative edits | PrunaAI | 3 credits | — | 16:9, 4:3, 1:1, 3:4, 9:16 |
2–10 February 2026
6 models · 3 providersKling, Qwen and Sourceful shipped 6 image models between 2 February 2026 and 10 February 2026. Of the 6, 3 have a spec page and 3 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Kling Image 3.0. Highest output in it: 4K on Riverflow 2.0.
- New name on the roster: Sourceful.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Riverflow 2.0Sourceful's Riverflow 2.0 Pro — production-grade image generation aimed at design and product work, where layout discipline and… | Sourceful | 11 credits | 4K | 1:1, 21:9, 16:9, 3:2, 4:3, 5:4 |
| Kling Image 3.0Kling V3 text-to-image generation with high-quality photorealistic and artistic outputs | Kling | 3 credits | 2K | 16:9, 9:16, 1:1, 4:3, 3:4, 3:2 |
| Qwen Image 2 EditQwen Image 2 Edit is an image editing model that allows users to modify and enhance images based on specific instructions. It is… | Qwen | 3 credits | — | 16:9, 4:3, 1:1, 3:4, 9:16 |
| Qwen Image 2 Edit ProvariantQwen Image 2 Edit Pro is an advanced image editing model that allows users to modify and enhance images based on specific… | Qwen | 6 credits | — | 16:9, 4:3, 1:1, 3:4, 9:16 |
| Qwen Image 2 Pro Text to ImagevariantQwen Image 2 Pro is a powerful text-to-image model that generates high-quality images from textual descriptions. It is designed… | Qwen | 6 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Qwen Image 2 Text to ImagevariantQwen Image 2 is a powerful text-to-image model that generates high-quality images from textual descriptions. It is designed to… | Qwen | 3 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
13–17 February 2026
7 models · 2 providersByteDance and Recraft shipped 7 image models between 13 February 2026 and 17 February 2026. Of the 7, 2 have a spec page and 5 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Seedream 5.0 Lite Edit. Highest output in it: 4K on Seedream 5.0 Lite.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Seedream 5.0 Lite EditvariantSeedream 5.0 for image editing | ByteDance | 3 credits | — | 16:9, 4:3, 1:1, 3:4, 9:16 |
| Seedream 5.0 LitevariantSeedream 5.0 Lite model for text to image generations | ByteDance | 3 credits | 4K | 1:1, 4:3, 3:4, 16:9, 9:16 |
| Recraft 4 Pro Text to ImagevariantRecraft 4 Pro is a powerful text-to-image model that generates high-quality images from textual descriptions. It is designed to… | Recraft | 20 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4 Pro Text to VectorvariantRecraft 4 Pro is a powerful text-to-image model that generates high-quality images from textual descriptions. It is designed to… | Recraft | 24 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4 Text to ImageRecraft 4 is a powerful text-to-image model that generates high-quality images from textual descriptions. It is designed to… | Recraft | 4 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4 Text to VectorvariantRecraft 4 is a powerful text-to-image model that generates high-quality images from textual descriptions. It is designed to… | Recraft | 7 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft V4Latest Recraft V4 text-to-image model with exceptional design quality and typography support | Recraft | 4 credits | — | 1:1, 4:3, 3:4, 3:2, 2:3, 16:9 |
26 February – 5 March 2026
3 models · 2 providersGoogle and Luma shipped 3 image models between 26 February 2026 and 5 March 2026. Of the 3, 2 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Nano Banana 2. Highest output in it: 4K on Nano Banana 2.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Nano Banana 2Ultra-fast lightweight image generation with Nano Banana 2 for rapid creative iteration | 3-6 credits | 4K | 8:1, 4:1, 21:9, 16:9, 3:2, 4:3 | |
| Nano Banana 2 EditvariantFast image editing model from Nano Banana 2 for quick prompt-based image modifications | 3-6 credits | 4k | 8:1, 4:1, 21:9, 16:9, 3:2, 4:3 | |
| Luma UNI 1 MaxLuma Uni-1 (Max) — a reasoning-first image model that plans intent before generating, tuned for art-directed, cinematic results… | Luma | 9 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
1 April 2026
3 models · 1 providerWan shipped 3 image models on 1 April 2026. Of the 3, 3 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Wan 2.7 Image to Image.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Wan 2.7 Image to ImagevariantWAN 2.7 transforms and edits existing images with text-guided instructions, supporting up to 4 reference images for creative… | Wan | 3 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Wan 2.7 Pro Text to ImagevariantWAN 2.7 Pro generates premium-quality images from text prompts with superior detail and composition | Wan | 6 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Wan 2.7 Text to ImagevariantWAN 2.7 generates high-quality images from text prompts with advanced prompt understanding and detailed output | Wan | 3 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
6–21 April 2026
4 models · 3 providersBaidu, OpenAI and Wan shipped 4 image models between 6 April 2026 and 21 April 2026. Of the 4, 3 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 1 credit on ERNIE Image Turbo. Highest output in it: 4K on GPT Image 2 Image to Image.
- New name on the roster: Baidu.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Wan 2.7 Pro EditWAN 2.7 Pro delivers professional-grade image editing and transformation with text-guided instructions, supporting up to 4… | Wan | 6 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| ERNIE Image TurboBaidu ERNIE Image (Turbo) — an 8B diffusion-transformer distilled for fast generation, strong at posters, comics, and… | Baidu | 1 credit | — | 1:1, 2:3, 3:2, 4:3, 3:4, 16:9 |
| GPT Image 2 Image to ImagevariantOpenAI GPT Image 2 image-to-image. Edits or remixes a reference image (up to 16 inputs) using a text prompt; supports up to 4K… | OpenAI | 9-17 credits | 4K | 3:1, 21:9, 2:1, 16:9, 3:2, 4:3 |
| GPT Image 2 Text to ImageOpenAI GPT Image 2 text-to-image. High-fidelity image generation with strong prompt adherence; supports up to 4K rendering. | OpenAI | 9-17 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
22 April – 8 May 2026
5 models · 2 providersHiDream and Imagine shipped 5 image models between 22 April 2026 and 8 May 2026. Of the 5, 4 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 1 credit on HiDream O1 Image. Highest output in it: 2K on ImagineArt 2.0.
- New name on the roster: HiDream.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| ImagineArt 2.0ImagineArt 2.0 text-to-image with true-to-life realism, precise text rendering, and cinematic lighting. Supports 1K/2K output and… | Imagine | 4 credits | 2K | 4:1, 3:1, 21:9, 16:9, 3:2, 4:3 |
| HiDream O1 ImageHiDream O1 pixel-native text-to-image that reasons before it draws. Unified model producing high-resolution images up to 2K, with… | HiDream | 1 credit | — | 1:1, 2:3, 3:2, 4:3, 3:4, 16:9 |
| HiDream O1 Image DevHiDream O1 Image (Dev) — the faster, distilled tier of the pixel-native text-to-image model. Lower cost with strong prompt… | HiDream | 1 credit | — | 1:1, 2:3, 3:2, 4:3, 3:4, 16:9 |
| HiDream O1 Image Dev EditvariantHiDream O1 image editing (Dev) — the distilled, faster editing tier. Instruction-driven edits on a supplied image at up to 2K. | HiDream | 1 credit | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| HiDream O1 Image EditHiDream O1 image editing. Instruction-driven edits and subject-driven changes on a supplied image at up to 2K resolution. | HiDream | 1 credit | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
12–14 May 2026
8 models · 2 providersKrea and Recraft shipped 8 image models between 12 May 2026 and 14 May 2026. Of the 8, 2 have a spec page and 6 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Krea 2 Medium.
- New name on the roster: Krea.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Krea 2 LargevariantKrea 2 Large text-to-image — the larger tier of Krea 2 with strong aesthetics, tunable creativity, and optional style references. | Krea | 5 credits | — | 1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5 |
| Krea 2 MediumKrea 2 Medium text-to-image. Fast, photoreal generation with tunable creativity levels and optional style references. | Krea | 3 credits | — | 1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5 |
| Recraft 4.1 Pro Text to ImagevariantRecraft V4.1 Pro text-to-image. Premium tier with enhanced detail, fidelity, and prompt adherence for professional use. | Recraft | 17 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4.1 Pro Text to VectorvariantRecraft V4.1 Pro text-to-vector. Premium tier vector graphics generation with crisper paths, richer detail, and better prompt… | Recraft | 24 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4.1 Text to ImageRecraft V4.1 standard text-to-image. High-quality image generation from prompts with strong style control and consistency. | Recraft | 3 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4.1 Text to VectorvariantRecraft V4.1 standard text-to-vector. Generates editable vector graphics (SVG) from text prompts — ideal for logos, icons, and… | Recraft | 7 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4.1 Utility Pro Text to ImagevariantRecraft V4.1 Utility Pro text-to-image. Pro-quality output via the utility SKU, priced the same as standard V4.1 Pro T2I. | Recraft | 17 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Recraft 4.1 Utility Text to ImagevariantRecraft V4.1 Utility tier text-to-image. Lightweight variant priced the same as standard V4.1 T2I. | Recraft | 3 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
1–2 June 2026
3 models · 2 providersMicrosoft and NVIDIA shipped 3 image models between 1 June 2026 and 2 June 2026. Of the 3, 2 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 4 credits on Cosmos 3 Super.
- New names on the roster: NVIDIA, Microsoft.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Cosmos 3 SuperNVIDIA Cosmos 3 Super text-to-image. High-fidelity generation with optional agentic refinement and prompt expansion for strong… | NVIDIA | 4 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Mai Image 2.5 EditMicrosoft MAI-Image-2.5 image editing (pixel-level, cleanup, backgrounds, text) | Microsoft | 4 credits | — | auto, 1:1, 4:3, 3:4, 16:9, 9:16 |
| Mai Image 2.5 Text to ImagevariantMicrosoft MAI-Image-2.5 photorealistic text-to-image with sharp text rendering | Microsoft | 4 credits | — | 16:9, 3:2, 4:3, 1:1, 3:4, 2:3 |
3–30 June 2026
3 models · 3 providersGoogle, Ideogram and Luma shipped 3 image models between 3 June 2026 and 30 June 2026. Of the 3, 1 has a spec page and 2 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 2 credits on Nano Banana 2 Lite.
- New name on the roster: Ideogram.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Ideogram V4Ideogram V4 text-to-image. Creates high-quality images, posters, and logos with crisp visuals, accurate text rendering, fine… | Ideogram | 5 credits | — | 1:1, 16:9, 9:16, 4:3, 3:4 |
| Luma UNI 1 Max EditvariantLuma Uni-1 Max Edit: reasoning-first image editing that plans the change before rendering. Rewrites only what the prompt asks… | Luma | 9 credits | — | — |
| Nano Banana 2 LitevariantGoogle Nano Banana 2 Lite - fastest, most cost-efficient Gemini image model, fixed 1K output, 14 aspect ratios | 2 credits | — | auto, 1:1, 2:3, 3:2, 3:4, 4:3 |
23 July – 1 August 2026
6 models · 2 providersMicrosoft and Qwen shipped 6 image models between 23 July 2026 and 1 August 2026. Of the 6, 1 has a spec page and 5 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Qwen Image 3 Image to Image. Highest output in it: 2K on Qwen Image 3 Image to Image.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| MAI Image 2.5 Pro EditvariantMicrosoft's flagship MAI-Image-2.5-Pro image editor: precise natural-language edits to one photo that keep composition, identity,… | Microsoft | 22 credits | — | auto, 1:1, 4:3, 3:4, 16:9, 9:16 |
| MAI Image 2.5 ProvariantMicrosoft's flagship MAI-Image-2.5-Pro text-to-image model: design-ready, highly detailed images with precise, legible typography… | Microsoft | 16 credits | — | auto, 1:1, 4:3, 3:4, 16:9, 9:16 |
| Qwen Image 3 Image to ImagevariantEdit or restyle up to three reference images with a natural-language instruction, preserving key details such as faces and… | Qwen | 3 credits | 2K | 1:1, 3:2, 2:3, 4:3, 3:4, 16:9 |
| Qwen Image 3 Pro Image to ImagevariantThe Pro tier of Qwen Image 3 editing — up to three reference images at higher quality, for finer detail and stronger identity… | Qwen | 6 credits | 2K | 1:1, 3:2, 2:3, 4:3, 3:4, 16:9 |
| Qwen Image 3 Pro Text to ImagevariantThe Pro tier of Qwen Image 3 — the same controls at higher quality, for finer detail and more reliable text rendering. | Qwen | 6 credits | 2K | 1:1, 3:2, 2:3, 4:3, 3:4, 16:9 |
| Qwen Image 3 Text to ImageQwen's third-generation image model. Strong prompt adherence and complex text rendering, in both Chinese and English, with… | Qwen | 3 credits | 2K | 1:1, 3:2, 2:3, 4:3, 3:4, 16:9 |
5–24 August 2026
4 models · 2 providersByteDance and Grok shipped 4 image models between 5 August 2026 and 24 August 2026. Of the 4, 2 have a spec page and 2 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 2 credits on Grok Imagine Image 2.0. Highest output in it: 2K on Seedream 5.0 Pro.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Seedream 5.0 ProByteDance's flagship text-to-image model — deep-thinking prompt understanding, native text rendering in 14 languages, and precise… | ByteDance | 6 credits | 2K | 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
| Seedream 5.0 Pro EditvariantGrounded, region-precise image editing — changes one element while keeping the rest of the frame intact, with layer separation,… | ByteDance | 6 credits | — | 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 |
| Grok Imagine Image 2.0xAI Grok Imagine 2.0 text-to-image. Thirteen aspect ratios from 2:1 through 1:2, 1K or 2K output, and up to four images per… | Grok | 2-3 credits | 2K | 2:1, 20:9, 19.5:9, 16:9, 4:3, 3:2 |
| Grok Imagine Image 2.0 EditvariantxAI Grok Imagine 2.0 image editing. Takes up to three source images and a text instruction; the default aspect ratio preserves… | Grok | 5-7 credits | 2K | auto, 2:1, 20:9, 19.5:9, 16:9, 4:3 |
8 September 2026
4 models · 1 providerOpenAI shipped 4 image models on 8 September 2026. Of the 4, 1 has a spec page and 3 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 4 credits on GPT Image 2.5 Flare Edit. Highest output in it: 4K on GPT Image 2.5 Flare Edit.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| GPT Image 2.5 Flare EditPrecise editing with OpenAI's default GPT Image 2.5 model: changes only what is asked, keeping subject, composition and… | OpenAI | 4-7 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
| GPT Image 2.5 Flare Text to ImagevariantOpenAI's default GPT Image 2.5 model. Fast, high-quality generation with natural lighting, rich textures and complex layouts,… | OpenAI | 4-7 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
| GPT Image 2.5 Sunburst EditvariantEditing with the tightest control from OpenAI's precision GPT Image 2.5 model: edits scoped exactly to the instruction, with… | OpenAI | 4-7 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
| GPT Image 2.5 Sunburst Text to ImagevariantOpenAI's precision-focused GPT Image 2.5 model, built for premium visual work with extra fidelity on intricate detail, in… | OpenAI | 4-7 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
23 September – 1 October 2026
4 models · 2 providersBlack Forest Labs and ByteDance shipped 4 image models between 23 September 2026 and 1 October 2026. Of the 4, 3 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Seedream 5.0 Flash. Highest output in it: 2K on Seedream 5.0 Flash.
- New name on the roster: Black Forest Labs.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Seedream 5.0 FlashSeedream 5.0 Flash is ByteDance Seed's fast, low-cost tier of the Seedream 5.0 image family (September 2026). It turns a text… | ByteDance | 3 credits | 2K | 1:1, 9:16, 16:9, 3:4, 4:3 |
| Seedream 5.0 Flash EditvariantSeedream 5.0 Flash Edit is the fast, low-cost editing tier of ByteDance Seed's Seedream 5.0 family. Describe the change and it… | ByteDance | 3 credits | 2K | 1:1, 9:16, 16:9, 3:4, 4:3 |
| FLUX 3 ImageFLUX 3 Image is Black Forest Labs' newest text-to-image model: it follows long, complex prompts closely and renders accurate text… | Black Forest Labs | 4 credits | 1K | 1:1, 21:9, 2:1, 16:9, 3:2, 7:5 |
| FLUX 3 Image EditFLUX 3 Image Edit from Black Forest Labs changes an image from a plain-text instruction and can combine up to 10 reference images… | Black Forest Labs | 4 credits | 1K | auto, 1:1, 21:9, 2:1, 16:9, 3:2 |
8–15 October 2026
6 models · 3 providersBlack Forest Labs, Google and Krea shipped 6 image models between 8 October 2026 and 15 October 2026. Of the 6, 3 have a spec page and 3 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 1 credit on FLUX 3 Image Klein. Highest output in it: 4K on Nano Banana 2.1.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Nano Banana 2.1Google's newest Nano Banana for image generation and editing, with improved image quality, prompt adherence, character… | 4-7 credits | 4K | 8:1, 4:1, 21:9, 16:9, 3:2, 4:3 | |
| FLUX 3 Image KleinFLUX 3 Image Klein is the small, fast model of Black Forest Labs' FLUX 3 family (October 2026): a 1K image in about five seconds… | Black Forest Labs | 1 credit | 1K | 1:1, 9:16, 16:9, 3:4, 4:3 |
| FLUX 3 Image Klein EditvariantFLUX 3 Image Klein Edit from Black Forest Labs changes an image from a written description of the result and can combine up to 4… | Black Forest Labs | 1 credit | 1K | 1:1, 9:16, 16:9, 3:4, 4:3 |
| Krea 3Krea 3 is Krea's newest image model (October 2026): photoreal images in seconds, with a built-in prompt planner that turns a… | Krea | 2 credits | 1K | 1:1, 9:16, 16:9, 3:4, 4:3 |
| Krea 3 EditvariantKrea 3 Edit changes a photo from a plain-text instruction in seconds while keeping products, characters and text consistent, and… | Krea | 4 credits | 1K | 1:1, 9:16, 16:9, 3:4, 4:3 |
| Krea 3 UltravariantKrea 3 Ultra is the highest-quality mode of Krea's October 2026 image model: it plans the edit, renders several candidates at 2K… | Krea | 28 credits | 2K | — |
19 October – 4 November 2026
6 models · 2 providersBria and OpenAI shipped 6 image models between 19 October 2026 and 4 November 2026. Of the 6, 2 have a spec page and 4 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 4 credits on Fibo 2.0. Highest output in it: 4K on GPT Image 2.6 Flare Edit.
- New name on the roster: Bria.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Fibo 2.0Fibo 2.0 is Bria's October 2026 image model, trained only on licensed data. It turns a written brief into a structured scene… | Bria | 4 credits | 1K | 1:1, 16:9, 9:16, 4:3, 3:4, 3:2 |
| Fibo 2.0 EditvariantFibo 2.0 Edit is Bria's licensed-data photo editor: describe the change in plain text and it restages the picture (a new… | Bria | 4 credits | 1K | — |
| GPT Image 2.6 Flare EditFast image editing with OpenAI's GPT Image 2.6 Flare: changes only what the prompt asks while keeping the subject, composition… | OpenAI | 7-11 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
| GPT Image 2.6 Flare Text to ImagevariantOpenAI's GPT Image 2.6 Flare, the fast default snapshot released November 2026: clean layouts, posters and product shots with… | OpenAI | 5-9 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
| GPT Image 2.6 Sunburst EditvariantPrecision editing with OpenAI's GPT Image 2.6 Sunburst: edits scoped tightly to the instruction, with subjects and composition… | OpenAI | 7-11 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
| GPT Image 2.6 Sunburst Text to ImagevariantOpenAI's GPT Image 2.6 Sunburst, the precision snapshot released November 2026, built for detailed compositions: dense scenes,… | OpenAI | 5-9 credits | 4K | auto, 1:1, 3:2, 2:3, 4:3, 3:4 |
9–10 November 2026
4 models · 2 providersByteDance and MiniMax shipped 4 image models between 9 November 2026 and 10 November 2026. Of the 4, 3 have a spec page and 1 is folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Seedream 5.5. Highest output in it: 4K on MiniMax H3 Image.
- New name on the roster: MiniMax.
| Model | Built by | Credits per job | Max output | Aspect ratios |
|---|---|---|---|---|
| Seedream 5.5Seedream 5.5 is ByteDance Seed's newest image model (November 2026). It turns a text prompt into a 2K picture with more natural… | ByteDance | 3 credits | 2K | 1:1, 4:3, 3:4, 16:9, 9:16, 2:3 |
| Seedream 5.5 EditvariantSeedream 5.5 Edit is the editing mode of ByteDance Seed's newest image model (November 2026). Describe a change and it edits your… | ByteDance | 5 credits | 2K | 1:1, 4:3, 3:4, 16:9, 9:16, 2:3 |
| MiniMax H3 Image EditMiniMax H3 Image Edit (November 2026) changes a photo from a plain-text instruction or combines up to 3 images into one 2K… | MiniMax | 7 credits | 2K | auto, 1:1, 16:9, 9:16, 4:3, 3:4 |
| MiniMax H3 ImageMiniMax H3 Image (November 2026) is MiniMax's open-weight image model built on its H3 video model. It turns a text prompt into a… | MiniMax | 3-7 credits | 4K | 1:1, 16:9, 9:16, 4:3, 3:4, 3:2 |
New AI image models by month in 2026
11 of the twelve months carried a image release; the busiest window was 12–14 May 2026, with 8. Each month links to the launch window where it starts.
Who shipped image in 2026
Recraft (11), OpenAI (10) and ByteDance (8) led on volume. SKUs, not quality — four tiers of one model count four times.
| Provider | Models | Spec pages | First | Latest |
|---|---|---|---|---|
| Recraft | 11 | 3 | 17 February 2026 | 14 May 2026 |
| OpenAI | 10 | 3 | 21 April 2026 | 4 November 2026 |
| ByteDance | 8 | 3 | 13 February 2026 | 9 November 2026 |
| Qwen | 8 | 2 | 10 February 2026 | 1 August 2026 |
| Krea | 5 | 2 | 12 May 2026 | 15 October 2026 |
| Black Forest Labs | 4 | 3 | 1 October 2026 | 15 October 2026 |
| Flux | 4 | 3 | 15 January 2026 | 15 January 2026 |
| 4 | 2 | 26 February 2026 | 8 October 2026 | |
| HiDream | 4 | 3 | 8 May 2026 | 8 May 2026 |
| Microsoft | 4 | 1 | 2 June 2026 | 23 July 2026 |
| Wan | 4 | 1 | 1 April 2026 | 6 April 2026 |
| Bria | 2 | 1 | 19 October 2026 | 19 October 2026 |
| Grok | 2 | 1 | 24 August 2026 | 24 August 2026 |
| Hunyuan | 2 | 1 | 26 January 2026 | 26 January 2026 |
| Luma | 2 | 1 | 5 March 2026 | 9 June 2026 |
| MiniMax | 2 | 2 | 10 November 2026 | 10 November 2026 |
| PrunaAI | 2 | 2 | 1 January 2026 | 1 February 2026 |
| Baidu | 1 | 1 | 15 April 2026 | 15 April 2026 |
| Ideogram | 1 | 1 | 3 June 2026 | 3 June 2026 |
| Imagine | 1 | 1 | 22 April 2026 | 22 April 2026 |
| Kling | 1 | 1 | 5 February 2026 | 5 February 2026 |
| NVIDIA | 1 | 1 | 1 June 2026 | 1 June 2026 |
| Sourceful | 1 | 1 | 2 February 2026 | 2 February 2026 |
What 2026 moved
Firsts, measured against every image model released before them. “New name on the roster” is the provider label in the catalog, not the company — one lab can hold several labels.
- 26 January – 1 February 2026
- New name on the roster: Hunyuan.
- 2–10 February 2026
- New name on the roster: Sourceful.
- 6–21 April 2026
- New name on the roster: Baidu.
- 22 April – 8 May 2026
- New name on the roster: HiDream.
- 12–14 May 2026
- New name on the roster: Krea.
- 1–2 June 2026
- New names on the roster: NVIDIA, Microsoft.
- 3–30 June 2026
- New name on the roster: Ideogram.
- 23 September – 1 October 2026
- New name on the roster: Black Forest Labs.
- 19 October – 4 November 2026
- New name on the roster: Bria.
- 9–10 November 2026
- New name on the roster: MiniMax.
Dates are each model’s released_at value, walked in order and grouped until a window held 3 or more. Credits are what one complete generation costs per the model’s own price matrix; where it states none, the headline rate is shown and the model sits out the cheapest-in-window line.
Other release years
Frequently asked questions
What are the newest AI image models?+
The newest image model in Versely's catalog is MiniMax H3 Image from MiniMax, released 10 November 2026. November 2026 is the latest month with image releases: 8 image models from OpenAI, ByteDance and MiniMax.
How many AI image models were released in 2026?+
Versely's catalog carries 84 image models with a 2026 release date, from 23 providers, arriving in 18 launch windows between 1 January 2026 and 10 November 2026.
What was the biggest AI image launch of 2026?+
12–14 May 2026, with 8 image models from Krea and Recraft. Of the 8, 2 have a spec page and 6 are folded into a parent model's page as tier or mode variants. Cheapest complete job in the window: 3 credits on Krea 2 Medium.
Which company released the most AI image models in 2026?+
Recraft, with 11 of the 84 image models dated 2026 — first on 17 February 2026, most recently on 14 May 2026. OpenAI shipped 10, ByteDance shipped 8, Qwen shipped 8.
What changed in AI image generation in 2026?+
Measured against everything the catalog carried before it: New name on the roster: Hunyuan; New name on the roster: Sourceful; New name on the roster: Baidu.
Run any 2026 image model in Versely
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync, in your browser or on your phone.
One account for the web and iPhone.