release

New models and developer products announced at DevDay

For AI workflow builders, these updates mean you can now process larger datasets cost-effectively and integrate advanced vision and image generation without managing separate infrastructure, enabling more sophisticated automation and user-facing features.

OpenAI Blog·November 6, 2023·1 min readrelease

releaseNew models and developer products announced at DevDay

openai.com

What happened

At OpenAI's DevDay, the company unveiled a suite of new models and developer tools aimed at reducing costs and expanding capabilities. Key announcements include GPT-4 Turbo, which offers a 128K context window—enough to process over 300 pages of text in a single prompt—at prices 2.75x lower than GPT-4 for input tokens and 2x lower for output tokens. The new Assistants API simplifies building agent-like experiences, allowing developers to create assistants with specific instructions, knowledge retrieval, and function calling. Additionally, GPT-4 Turbo with Vision brings multimodal understanding, while the DALL·E 3 API enables programmatic image generation. These updates represent a significant shift toward more accessible and powerful AI tools for developers, reducing the barrier to entry for complex applications.

Key takeaways

GPT-4 Turbo features a 128K context window and reduced pricing (input: $0.01 per 1K tokens, output: $0.03 per 1K tokens).
The Assistants API provides a managed environment for building custom AI assistants with persistent threads and built-in code interpreter.
GPT-4 Turbo with Vision accepts image inputs for analysis alongside text prompts.
DALL·E 3 is now available via API for direct integration into applications.