ChatGPT’s Image Generation Now Accessible via API by OpenAI
Integration of Image Generation in Businesses
People are now able to incorporate images inspired by Studio Ghibli, created by AI through ChatGPT, directly into their business operations. OpenAI has announced the addition of the model that powers its popular image generation tool to its API, earlier known for its use in ChatGPT.
Introducing the gpt-image-1 Model
The gpt-image-1 model provides developers and businesses with the capability to seamlessly integrate high-quality, professional-grade image creation into their own platforms and tools. OpenAI emphasizes that this model can produce images in various styles, adhere to specific guidelines, utilize vast world knowledge, and accurately render text. This versatility opens doors to countless practical uses across multiple sectors.
Pricing Details
The pricing for OpenAI’s API is structured around tokens for both text and image generation:
- Text input tokens: $5 per million tokens.
- Image input tokens: $10 per million tokens.
- Image output tokens: $40 per million tokens.
In comparison, competitors such as Stability AI and Google have different pricing structures. Stability AI operates on a credit-based system, with one credit amounting to $0.01, while Google’s Imagen model charges users $0.03 per generated image through the Gemini API.
Image Generation Feature Within ChatGPT
In April, OpenAI introduced an image generation and editing feature in the ChatGPT interface, making it more interactive. According to OpenAI, this new feature quickly gained immense popularity, with over 130 million users generating 700 million images within the first week of its introduction. Social media buzzed with users generating images themed around Studio Ghibli, which led OpenAI’s CEO, Sam Altman, to humorously remark that their GPUs were “melting” due to the excessive use.
Enterprise Applications of the Model
Businesses are seeking efficient ways to produce images for their projects without needing to switch between applications. OpenAI’s gpt-image-1 model allows companies to integrate this feature directly into their workflow ecosystems. The versatility of this model has already attracted interest from several enterprises and startups, with OpenAI citing various well-known brands using it for creative purposes.
Notable Integrations
Some companies exploring the potential of gpt-image-1 include:
- Canva, which is looking into ways to incorporate the model into its AI and Magic Studio Tools.
- GoDaddy, which has begun testing image generation for logo creation for its customers.
- Airtable, facilitating marketing and creative teams in managing asset workflows more effectively.
Safety Measures and Data Privacy
OpenAI has ensured that the API includes safety measures similar to those found within ChatGPT. Images generated by gpt-image-1 will include metadata from the Coalition for Content Provenance and Authenticity (C2PA), which categorizes content as AI-generated and tracks ownership. OpenAI actively participates in C2PA’s steering committee.
Furthermore, users will have the ability to manage content moderation to create images that align with their brand identity. Importantly, OpenAI has committed to not utilizing customer API data, including images generated or uploaded, to train its models.
