The article explains how the launch of GPT-4o mini, a powerful AI model from OpenAI, will revolutilize the AI business by offering a variety of applications.
This AI model from OpenAI, the market leader in the AI industry, is beneficial for tasks requiring the chaining or parallelization of multiple model calls, managing extensive context inputs, and providing rapid real-time responses in customer interactions. It is expected that lowering of the price for AI solutions will enhance the acceptance among fans, who are yet to experiment costly AI software.
Key Features of GPT-4o Mini:
Multimodal Support: GPT-4o mini currently supports text and vision in the API. Future updates will extend its capabilities to handle image, video, and audio inputs and outputs.
Expanded Context Window: With a context window of 128K tokens, it supports up to 16K output tokens per request.
Advanced Knowledge Base: The model’s knowledge is updated until October 2023.
Superior Performance: GPT-4o mini surpasses GPT-3.5 Turbo and other small models in both textual intelligence and multimodal reasoning.
Benchmark Performance:
Reasoning Tasks: GPT-4o mini excels in reasoning tasks involving both text and vision, achieving 82.0 percent on the MMLU benchmark, compared to 77.9 percent for Gemini Flash and 73.8 percent for Claude Haiku.
Math and Coding Proficiency: The model demonstrates exceptional mathematical reasoning and coding skills, scoring 87.0 percent on MGSM (math reasoning) and 87.2 percent on HumanEval (coding performance). These scores significantly outperform those of Gemini Flash and Claude Haiku.
Multimodal Reasoning: On the MMMU benchmark for multimodal reasoning, GPT-4o mini scored 59.4 percent, higher than Gemini Flash’s 56.1 percent and Claude Haiku’s 50.2 percent.
Availability and Pricing:
GPT-4o mini is now available in the Assistants API, Chat Completions API, and Batch API. The pricing is set at 15 cents per 1 million input tokens and 60 cents per 1 million output tokens, which equates to roughly 2500 pages in a standard book.
Rollout Plan:
ChatGPT Users: Free, Plus, and Team users can access GPT-4o mini starting today, replacing GPT-3.5.
Enterprise Users: Access will be granted starting next week.
Fine-Tuning: OpenAI plans to introduce fine-tuning capabilities for GPT-4o mini in the near future.
With its robust capabilities and cost-effective pricing, GPT-4o mini is set to be a game-changer in the AI landscape, offering enhanced performance for a wide range of applications.
