Provider
OpenAI AI Models on Martini
Martini hosts OpenAI image models GPT Image and GPT Image 2 alongside Sora 2 video, so storyboard frames, text-rich product shots, and cinematic clips run on the same canvas. Chain image generations into video, swap models per shot, and pay credit-by-credit instead of juggling separate OpenAI subscriptions.
About OpenAI
OpenAI ships two model families on Martini: GPT Image (the original DALL-E successor used inside ChatGPT) and GPT Image 2, the upgraded version with sharper text rendering, better prompt adherence, and stronger photoreal output. They sit beside Sora 2, OpenAI's flagship video model known for physics-aware motion, native audio, and multi-shot storyboard control. All three are exposed as Martini nodes so you can mix them with any other provider on the same canvas instead of context-switching between ChatGPT, the OpenAI API, and Sora.app.
On Martini the OpenAI lineup is positioned as the strongest pair for narrative video work. GPT Image 2 is the model most teams reach for when they need legible signage, UI mockups, packaging copy, or any image where typography matters more than painterly style. Sora 2 then takes those frames and produces 1080p clips with consistent characters across cuts. Because both models live in the same project, you can fan out a six-shot storyboard from one prompt, animate each frame, and stitch the result without leaving the canvas.
Pricing on Martini is metered per generation and shares a single wallet with every other provider, which is the main reason teams move OpenAI work here from a stack of separate subscriptions. Credits roll over from your monthly plan into wallet top-ups, and workspace billing keeps OpenAI usage attributable per member for clean client invoicing. If a request hits OpenAI rate limits or content policy filters, the same canvas lets you fail over to Imagen 4, Nano Banana 2, or Veo without rebuilding the workflow from scratch.
Available OpenAI models on Martini
Image models
Best use cases
- Storyboard frames where each shot needs legible captions, UI text, or product labels
- Product visuals with packaging copy, pricing, or brand typography baked into the image
- Multi-shot narrative video that needs consistent characters and physics-aware motion
- Cinematic short-form content where audio and visuals are generated together
- Concept art for ads where the prompt-to-image translation has to be highly literal
- Image-to-video chains that start in GPT Image 2 and finish in Sora 2 for a polished hero clip
Recommended workflows
AI Storyboard Generator
GPT Image 2 nails the typography and composition required for storyboard frames before Sora 2 animates them.
AI Image to Video
Generate a still in GPT Image 2 and pipe it directly into Sora 2 for a cinematic 1080p clip.
Multi-Shot AI Video
Sora 2 storyboard mode keeps characters and lighting consistent across multi-shot sequences.
AI Product Video Generator
Pair GPT Image 2 packaging shots with Sora 2 motion to build product video from a single brief.
Related how-to guides
Related features
Related reading
Other providers
Frequently asked questions
Do I need an OpenAI API key to use GPT Image or Sora 2 on Martini?
No. Martini provides hosted access to GPT Image, GPT Image 2, and Sora 2 — you pay with Martini credits and never touch an OpenAI key.
Is Sora 2 on Martini the same model as Sora.app?
Yes, Sora 2 on Martini is the official OpenAI Sora 2 model, exposed through their API so the underlying generation matches what Sora.app and ChatGPT serve.
When should I pick GPT Image 2 over Nano Banana 2 or Imagen 4?
Choose GPT Image 2 when typography, UI text, or packaging copy must read cleanly. Imagen 4 wins for cinematic photoreal lighting and Nano Banana 2 wins for fast iteration and reference-driven character consistency.
Can I chain GPT Image 2 into Sora 2 in one workflow?
Yes — drop a GPT Image 2 image node, connect its output to a Sora 2 video node, and Sora 2 will use the still as the opening frame for image-to-video.
Does Sora 2 generate audio?
Sora 2 supports native audio generation including ambient sound and dialogue, so you can ship clips without a separate voiceover or music pass for many use cases.
Are OpenAI requests on Martini subject to OpenAI content policy?
Yes, OpenAI moderation applies to GPT Image, GPT Image 2, and Sora 2 generations. If a prompt is blocked you can usually run the same idea through Imagen 4 or Seedance 2 from the same canvas.
Build with OpenAI on Martini
Open the canvas and wire OpenAI into the rest of your stack in minutes. Free to start — no card required.