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.
Read more: https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/