8 New AI Models Now on Nano Banana: Flux 2 Pro, Qwen Image, Ideogram V3, and More (2026)
We've added 8 new AI image and video models in the past two months. Flux 2 Pro for premium stills, Qwen Image for multilingual prompts, Ideogram V3 for text-in-image, plus Kling 3.0, Wan 2.7, Hailuo 2.3, and Seedance 2 Mini for video.
We've been busy. Over the past two months, Nano Banana has added 8 new AI models to the platform — covering both image generation and video generation from providers like Black Forest Labs, Alibaba, Ideogram, Kuaishou, MiniMax, and ByteDance. All available in one place, no switching between tools.
New Image Models
Flux 2 Pro — Black Forest Labs
The flagship model from the creators of Stable Diffusion. Flux 2 Pro delivers crisp detail, strong typography, and multi-image reference support — upload up to 4 reference images for precise control.
- Best for: Premium product shots, editorial images, poster design
- Resolutions: 1K (5 credits), 2K (10 credits)
- Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3
Qwen Image — Alibaba
Alibaba's budget-friendly image model with strong multilingual text understanding. If you write prompts in Chinese, Japanese, or Arabic, Qwen handles them better than most English-first models.
- Best for: Budget generation, multilingual prompts, quick edits
- Resolution: Auto (5 credits)
- Aspect ratios: 1:1, 4:3, 3:4, 16:9, 9:16
Ideogram V3 — Ideogram
The best text-in-image model on the market. Need readable text inside a poster, logo, or banner? Ideogram V3 delivers crisp, accurate typography that other models struggle with.
- Two tiers: Turbo (5 credits, fast) and Quality (10 credits, premium)
- Best for: Posters, logos, banners, any image with text
- Text-to-image only
New Video Models
Kling 3.0 Series — Kuaishou
Three variants under one family:
- Kling 3.0 Turbo — Fast generation, 720p/1080p, 5-10s. Best for quick results.
- Kling 3.0 — Full quality with optional native audio. 720p/1080p, 5-10s.
- Kling 3.0 Motion Control — New capability. Upload a reference photo and a driving video, and the AI transfers the motion onto your character. Read the full Motion Control guide →
Wan 2.7 — Alibaba
Alibaba's cinematic video model with text-to-video, image-to-video, and reference-to-video modes. Up to 1080p, 15-second clips with strong prompt following.
- Best for: Cinematic scenes, multi-reference video generation
- Resolutions: 720p, 1080p | Duration: 3-15s
Hailuo 2.3 / 2.3 Pro — MiniMax
MiniMax's latest image-to-video models. Hailuo 2.3 Pro adds top-tier detail and cinematic dynamics. Both support 6s and 10s durations at 768P/1080P.
- Best for: Animating still photos, cinematic short clips
- Hailuo 2.3: 25-40 credits | Pro: 40-75 credits
Seedance 2 Mini — ByteDance
ByteDance's budget video model. Text-to-video, image-to-video, and reference-to-video at 480p/720p, 5-10s. The most affordable way to generate video with audio.
- Best for: Budget video generation, quick tests
- Pricing: From 45 credits (480p-5s) to 190 credits (720p-10s)
Model Comparison at a Glance
| Model | Provider | Type | Starting Credits |
|---|---|---|---|
| Flux 2 Pro | Black Forest Labs | Image (t2i/i2i) | 5 |
| Qwen Image | Alibaba | Image (t2i/i2i) | 5 |
| Ideogram V3 | Ideogram | Image (t2i) | 5 |
| Kling 3.0 | Kuaishou | Video (t2v/i2i/v2v) | 65 |
| Wan 2.7 | Alibaba | Video (t2v/i2v/r2v) | 75 |
| Hailuo 2.3 | MiniMax | Video (i2v) | 25 |
| Seedance 2 Mini | ByteDance | Video (t2v/i2v/r2v) | 45 |