OpenAI has launched GPT-4o mini, a cost-efficient small language model designed to enhance accessibility and affordability for various applications.

Key Highlights and Features of GPT-4o Mini

1. Performance

GPT-4o mini outperforms GPT-4 in chat preferences, as reflected in the LMSYS leaderboard.
It achieves an impressive score of 82% on the MMLU benchmark.

2. Cost-Efficiency

The model is priced at 15 cents per million input tokens and 60 cents per million output tokens.
This pricing makes it significantly cheaper than previous models, including being 60% less expensive than GPT-3.5 Turbo.

3. Capabilities

GPT-4o mini supports both text and vision in its API.
It features a large context window of 128K tokens and includes knowledge up to October 2023.

4. Benchmark Performance

The model surpasses GPT-3.5 Turbo and other small models on various academic benchmarks, excelling in textual intelligence and multimodal reasoning.

5. Safety Measures

Built-in safety features help filter out undesirable content.
The model uses reinforcement learning with human feedback (RLHF) to align with OpenAI's safety policies.

6. Availability

GPT-4o mini is available through the Assistants API, Chat Completions API, and Batch API.
It is accessible to ChatGPT Free, Plus, and Team users, replacing GPT-3.5.

7. Future Developments

OpenAI plans to continue enhancing model capabilities while further reducing costs, aiming for seamless integration of AI into applications and websites.

This launch marks a significant advancement in making powerful AI models more accessible and affordable for developers and users.