Images, video, music and AI chat in one studio
Nano Banana Pro, GPT Image, Seedream, Kling, Claude, GPT‑6 and Suno in one place. Buy a token pack and spend it on anything in the studio, no subscription. The studio runs in your browser.
2 free tokens to start — enough for a couple of images, AI chat answers or song lyrics.
Images
Create from a text description or edit your own photos: swap the background, remove what is in the way, shoot a product for a marketplace, make a cover or an avatar. References help keep a style, a character or a product consistent from frame to frame.
Nano Banana 2 · Nano Banana Pro · Flux 2 · GPT Image 2 · Grok Imagine
All image models →
Video
Bring a still to life or make a clip from a description — for stories, ads and music videos. Some models take a first and a last frame, some can add sound. You choose the length and the resolution, and you see the price before you start.
Kling 2.6 · Seedance 2.0 · Kling 3.0 · Seedance 2.5 · Seedance 2.0 Mini
All video models →AI chat
Claude, GPT‑6, Gemini, Grok and DeepSeek in one service. Ask it to come up with names or a plan for a post, go through a document or a photo, write a text — and draw an image for it right in the conversation. Each answer is billed by what it actually used, usually a fraction of a token.
Claude Opus 5.5 · Claude Sonnet 5.5 · GPT-6 Sol · GPT-6 Luna · Gemini 3.8 Flash · Grok 4.7 · DeepSeek V4.1 Flash
Chat models and prices →Come up with three names for a seaside coffee shop and draw a sign for the best one
“Breeze”, “Gull & Bean”, “Fifth Wave”. I like the second one best — here is the sign:
Music
A song from a description or your own lyrics, a cover in another style, a continuation of a track, your own recording split into stems, and MIDI. Save a singing manner as a persona or create your own voice.
Suno V6 · Suno V6 Mini · Suno V6 Wild
Operations and prices →- Night Tram synthwave, vocals
- Balcony Morning lo-fi, piano
- Vocals, drums, bass your own track split into stems
Models and prices
The cost of one generation in tokens.
Images
Price of one image in tokens
- Nano Banana 2 The go-to model for most tasks — posts, illustrations, images with crisp in-image text, up to 14 reference images per request from 2 t
- Nano Banana Pro For complex compositions, infographics, and precise in-image text at 2K/4K — final assets for advertising and print from 5 t
- Flux 2 Photorealistic product shots, ad scenes, and precise edits to existing images without redrawing from scratch from 2 t
- GPT Image 2 Realistic photos, packaging, and UI mockups with precise typography and dense in-image text 3 t
- Grok Imagine A distinctive artistic look for covers, social posts, and creative ideas — when you want to break out of the default style 2 t
Video
Price of one clip in tokens
- Kling 2.6 5–10 second videos with speech, effects, and ambient sound — for ads and social media from 8 t
- Seedance 2.0 Cinematic video up to 15 seconds in 1080p — natural motion physics and synced audio from 16 t
- Kling 3.0 3–15 second video in 720p and 1080p with native audio — plus fast Turbo and reference-driven Omni from 12 t
- Seedance 2.5 Video up to 30 seconds in 1080p — longer scenes, precise reference control, and synced audio from 28 t
- Seedance 2.0 Mini Fast, low-cost video up to 15 seconds — Seedance 2.0 Fast–level quality for high-volume work from 4 t
AI chat
Approximate price of one answer in tokens
- Claude Opus 5.5 The smartest: hard tasks, code, long texts 0.3 t
- Claude Sonnet 5.5 Smart and fast: writing, code, analysis 0.2 t
- GPT-6 Sol OpenAI's strong all-round model 0.1 t
- GPT-6 Luna Fast and very cheap 0.1 t
- Gemini 3.8 Flash Google's fast model 0.1 t
- Grok 4.7 xAI's model: reasoning and code 0.1 t
Music
Price of one operation in tokens
- Song Two versions of a song from a description or from your own lyrics and style. 3 t
- Cover in another style The same melody in a new style, two versions. 3 t
- Extend Continues the track from the chosen second, two versions. 3 t
- Split your track Your recording split into stems: vocals and instrumental, all instruments, or one chosen stem. File up to 20 MB, it will appear in the library. from 2.5 t
- Sounds and loops Sound effects, ambience and loops from a description. 0.7 t
Token packs
You see the price of a generation before you run it.
-
Start
$4.99
439 KGS
30 tokens
$0.166 per token
15 KGS per token
Buy -
Popular
$9.99
878 KGS
80 tokens
$0.125 per token · −30%
11 KGS per token
Buy -
Pro
$29
2 549 KGS
300 tokens
$0.097 per token · −42%
8.5 KGS per token
Buy
Prices are in US dollars and Kyrgyz som.
Ready-made templates
Pick an example, drop in your photo or text — the prompt is already written.
-
Instant photoYour photo as a 90s snapshot
-
Product photoYour product on a podium in soft studio light
-
AvatarA hand-drawn portrait
-
Track coverFor a track, a podcast or a video
-
Boxed figureYou as a collectible toy
-
Greeting cardA card for any holiday
-
ClayA character in plasticine style
-
GlossyA photo in fashion-magazine style
Questions
How do I pay?
With VISA cards and Kyrgyz banks, or with cryptocurrency: USDT, BTC, ETH and other popular coins. Packs are bought once, with no recurring charges.
What are tokens?
The studio's internal currency. Every model and operation has its own price in tokens — you see it in the catalogue and before you run. An answer in the chat costs what it actually used, usually a fraction of a token.
Can I try it for free?
Yes. A new user gets 2 free tokens. You can use them to try creating an image or chatting with AI. Until you top up, the first images carry a watermark.
What happens if a generation fails?
If the model doesn't return a result, no tokens are charged. This applies to images, video, music and AI chat.
Can I use the results commercially?
Yes. Rights to the result follow the terms of the selected model, and most modern models allow commercial use.
Do you have an MCP server? I want to connect NanoBananana to my AI agent.
Yes. The MCP server is ready and in closed testing, and it will be available to everyone soon.
Didn't find an answer? Write to support@kosmosoid.dev.