Google AI Studio
Browser workspace to prototype with Gemini, Veo, Nano Banana and Lyria, then build and export an app
Google AI Studio is Google's browser workspace for testing Gemini and media models (Nano Banana image models, Veo 3.1 and Gemini Omni Flash video, Lyria 3.5 music, Gemini TTS voices) and for building apps with the Build agent. Prototypes can be exported to Google Antigravity with the full agent conversation. It is a development surface, not a generator in its own right: the models behind it are billed per token or per output through the Gemini API.
- 14 creators using
- Freemium pricing
- Easy learning curve
What Google AI Studio can do
- Infinite canvas No
- AI agent Partial
- Node / flow builder No
- Public API Yes
- MCP server Not confirmed
14 of 26 capabilities confirmed
| Capability | Supported | Details |
|---|---|---|
| Text to image | Yes | Nano Banana 2.1, Nano Banana 2, Nano Banana Pro |
| Image editing | Yes | Nano Banana 2.1 generation and editing |
| Text to video | Yes | Veo 3.1 family and Gemini Omni Flash |
| Image to video | Yes | Gemini Omni Flash first-and-last-frame interpolation |
| Video to video | Partial | Omni Flash video extension and conversational edits |
| Native audio in video | Yes | Veo 3.1 Standard generates video with audio |
| Lip-sync | Not confirmed | |
| Voice / TTS | Yes | Gemini 3.8 Flash TTS, 150+ voices, voice design and replication |
| Music & SFX | Yes | Lyria 3.5 full-length songs |
| Upscaling | Not confirmed | |
| Character consistency | Not confirmed | |
| Camera control | Not confirmed | |
| Custom training | Not confirmed | |
| Third-party models | No | Gemini-family models only |
| Capability | Supported | Details |
|---|---|---|
| Infinite canvas | No | |
| Node / flow builder | No | |
| AI agent | Partial | Build agent (Antigravity harness) plans and builds apps from a prompt; it does not storyboard or generate a film |
| Chat assistant | Yes | Chat prompt playground and Build agent chat |
| Team collaboration | Not confirmed |
| Capability | Supported | Details |
|---|---|---|
| Public API | Yes | Gemini API, keys created in AI Studio |
| MCP server | Not confirmed | |
| Mobile app | Partial | |
| Desktop / local | Partial | Export to Antigravity desktop app for local development |
| Open source | No |
| Capability | Supported | Details |
|---|---|---|
| Free tier | Yes | Some models free with limited access; free tier requires an active project or trial |
| Commercial use | Not confirmed |
Yes Partial No Not confirmed by the vendor
What makes Google AI Studio different
- 01
Same models as the Gemini API with a free tier for testing: Veo 3.1, Omni Flash, Nano Banana and Lyria are all callable from one browser workspace before any billing is set up.
- 02
Native audio on video: Veo 3.1 Standard generates video with synced audio at $0.40/s for 720p or 1080p, which most rivals in this catalog bill separately.
- 03
Build agent to Antigravity handoff: a prototype moves to local development with code and the agent conversation, instead of a one-way code copy.
- 04
Multimodal input breadth: Gemini 3.8 Flash accepts text, image, video, audio and PDF in one request with a 1,048,576-token context, useful for reviewing a full cut against a script.
- 05
Clear per-token and per-second pricing on one public page, with rate-limit tiers that step up automatically after spend.
About Google AI Studio
Google AI Studio is the browser entry point to the Gemini API. You can prompt models in chat, adjust settings, generate images, video, speech and music from the same interface, and create API keys. Its Build mode is an agent that turns a prompt into a working app, and since Google I/O 2026 (May) Build runs on the Antigravity agent harness.
What changed in 2026: Gemini 3.8 Flash reached GA on Sep 2, Gemini Omni Flash (video generation and conversational video editing, with video extension and first-and-last-frame interpolation) went GA on Aug 27, Nano Banana 2.1 went GA on Oct 6, Gemini 3.8 Flash TTS and Lyria 3.5 (full-length songs) arrived in September. Build prototypes can now be exported to Antigravity with code and agent history.
For filmmakers, AI Studio is most useful as a cheap place to test shot prompts, storyboard frames and image edits against the same Veo, Omni and Nano Banana models that power Google's other creative apps, and to build small internal tools (shot lists, contact-sheet viewers, batch renamers) without leaving Google's stack. It is not a timeline, a canvas or a production suite; video and image outputs need to be downloaded and finished elsewhere.
What's new in Google AI Studio
-
Nano Banana 2.1 GA
Image generation and editing model, about $0.034 per 1K image on the paid tier; Gemini 3.1 Flash Image is marked deprecated.
Source: ai.google.dev ↗ -
Gemini 3.8 Flash and 3.8 Flash TTS GA
Gemini 3.8 Flash went GA on Sep 2 with a 1M-token context; Flash TTS GA on Sep 22 adds voice design, voice replication and a 150+ voice library.
Source: ai.google.dev ↗ -
Lyria 3.5 music GA
Full-length song generation with 44.1 kHz stereo output, available through the Gemini API.
Source: ai.google.dev ↗ -
Gemini Omni Flash GA
Video generation and conversational editing with video extension and first-and-last-frame interpolation; the preview model shuts down Sep 30, 2026.
Source: ai.google.dev ↗ -
Build mode moves to the Antigravity harness
At Google I/O, AI Studio Build switched to the Antigravity agent harness and gained an export to Antigravity that carries code and full agent conversation.
Source: antigravity.google ↗
Models available in Google AI Studio
-
Veo 3.1 Fast Latest $$$ Google Fast $0.10/s at 720p; Lite $0.05/s at 720p, no 4K Also: Veo 3.1 Lite · Veo 3.1 Standard
-
-
-
-
-
$ · $$ · $$$ = relative cost against other models of the same type, on fal.ai list prices. Cost inside Google AI Studio depends on its own plans and credits.
Key Features
- Prompt and chat playground for Gemini models with adjustable settings
- Image generation and editing with Nano Banana 2.1, Nano Banana 2 and Nano Banana Pro
- Video generation with Veo 3.1 (synced audio) and Gemini Omni Flash (extension, first-and-last frame)
- Speech generation with Gemini 3.8 Flash TTS voice library, voice design and replication
- Music generation with Lyria 3.5
- Build agent that turns a prompt into a working app, running on the Antigravity harness
- Export a Build prototype to Antigravity with code and agent conversation
- API key creation for the Gemini API
- Usage logs for Interactions API calls in the AI Studio dashboard
When to reach for it — and when to skip
Reach for it when…
- Free tier lets you test Veo, Omni and Nano Banana output before paying
- Native audio with Veo 3.1 cuts a separate dialogue or SFX step
- Clear per-second and per-image prices on one public pricing page
- Same API powers prototype and production, so a test prompt ports directly
Skip it when…
- Gemini-family models only; no access to Kling, Runway or Seedance from here
- Several 3.x rates are promotional until Dec 31, 2026, then roughly double
- Free-tier inputs may be used to improve Google products
- Video outputs are not edited in place; you need a separate NLE or editor afterwards
- Model IDs are retired often (Gemini 2.5 Flash Image shuts down Oct 2, 2026; Omni preview Sep 30, 2026)
Best For
✓ Ideal for
- Testing shot prompts and image edits on Nano Banana before committing to a paid render budget
- Generating short Veo 3.1 clips with synced dialogue or ambience for pre-viz
- Building a small internal tool for a shoot, such as a shot-list tracker or a contact-sheet viewer
- Generating voiceover drafts with Gemini TTS for a rough cut
✗ Not built for
- Multi-model storyboarding across vendors: use an aggregator or a node tool instead
- Finishing edits, grading or timeline work
Working Tips from Filmmakers Using Google AI Studio
- 01 Run the same shot prompt on Veo 3.1 Fast ($0.10/s at 720p) first, then re-render the keeper on Standard ($0.40/s) to cut test spend by about 75%
- 02 Generate stills in Nano Banana 2.1 at 1K for exploration ($0.034) and only upscale the chosen frame to 2K or 4K
- 03 Use the Batch API for overnight renders of prompt variants; it is billed at 50% of the standard rate
- 04 Set billing and a spend cap before a long test session, because the free tier has no spend ceiling for paid-only features
Google AI Studio pricing in 2026
| Plan | Price | What you get |
|---|---|---|
| Free | $0 none |
|
| Paid (pay-as-you-go) | Per token or per output unit usage |
|
| Enterprise | Custom contract |
|
Prices checked against Google AI Studio's pricing page on 9 Oct 2026. Vendors change plans often, so confirm before you buy.
Google AI Studio vs Replicate vs Claude
| Google AI Studio | Replicate | Claude | |
|---|---|---|---|
| Price | Free tier for some models; paid is pay-as-you-go per token or per output (Veo 3.1 Standard from $0.40/s) | Pay per second of hardware: $0.000225/s (T4) to $0.001525/s (H100). No monthly plan on the pricing page; enterprise is custom. | Free; Pro $20/mo ($17/mo billed annually); Max from $100/mo |
| Infinite canvas | No | No | Partial |
| AI agent | Partial | No | Yes |
| Chat assistant | Yes | No | Yes |
| Team collaboration | Not confirmed | Not confirmed | Yes |
| Text to image | Yes | Yes | No |
| Image editing | Yes | Not confirmed | No |
| Text to video | Yes | Yes | No |
| Image to video | Yes | Yes | No |
| Video to video | Partial | Not confirmed | No |
| Native audio in video | Yes | Yes | No |
| Voice / TTS | Yes | Yes | No |
| Music & SFX | Yes | Yes | No |
| Custom training | Not confirmed | Yes | No |
| Third-party models | No | Yes | No |
| Public API | Yes | Yes | Yes |
| MCP server | Not confirmed | Yes | Partial |
| Mobile app | Partial | Not confirmed | Yes |
| Desktop / local | Partial | Partial | Yes |
| Open source | No | Partial | No |
| Free tier | Yes | Not confirmed | Yes |
| Commercial use | Not confirmed | Not confirmed | Yes |
Each column comes from that tool's own researched page. "Not confirmed" means the vendor doesn't say.
Open the full Google AI Studio vs Replicate vs Claude comparison →Use Cases
Tags
Creators Using Google AI Studio
Sources
Facts on this page were checked against Google AI Studio's own pricing, product and changelog pages on , then cross-checked with independent coverage where vendor pages were silent. Anything we couldn't confirm is marked "Not confirmed" rather than guessed.
- Gemini API pricing (updated 2026-10-07) ai.google.dev
- Gemini API release notes, Jun-Oct 2026 ai.google.dev
- Rate limits and usage tiers ai.google.dev
- Gemini 3.8 Flash model page ai.google.dev
- Google I/O 2026 Antigravity post (Build mode and export) antigravity.google
- Google AI Studio home aistudio.google.com
Spotted something out of date? Tell us and we'll re-check it.
Discussion
No comments yet — be the first.