OpenAI launches cost-effective GPT-4o mini

OpenAI unveiled its latest small AI model, GPT-4o mini, on Thursday, positioning it as a faster and more cost-effective alternative to its current cutting-edge AI offerings.

The model is now available for developers via OpenAI’s API and through the ChatGPT web and mobile app for consumers starting today, with enterprise access slated for next week.

According to OpenAI, GPT-4o mini excels in reasoning tasks involving text and vision, outperforming leading small AI models in the industry. These smaller models are gaining popularity among developers for their efficiency in handling high-volume, repetitive tasks compared to larger counterparts like GPT-4 Omni or Claude 3.5 Sonnet.

Read Also:

Bitcoin, Ethereum lead $17.8B crypto institutional investment

TikTok partners DistroKid to speed up artist account creation

Replacing GPT-3.5 Turbo as OpenAI’s smallest model, GPT-4o mini achieves an 82% score on the MMLU benchmark for reasoning, surpassing Gemini 1.5 Flash at 79% and Claude 3 Haiku at 75%, as reported by Artificial Analysis.

Cost-efficiency is a significant highlight, with GPT-4o mini being more than 60% cheaper to operate than its predecessor. Initially supporting text and vision in its API, the model will expand to include video and audio capabilities in the future.

Olivier Godemont, OpenAI’s Head of Product API, emphasized the model’s affordability: “For every corner of the world to be empowered by AI, we need to make the models much more affordable. I think GPT-4o mini is a really big step forward in that direction.”

Developers can access the AI model through OpenAI’s API, priced at 15 cents per million input tokens and 60 cents per million output tokens. The model features a context window of 128,000 tokens and has a knowledge cutoff of October 2023.

While specifics on its size were not disclosed, OpenAI compared the AI model to other small AI models like Llama 3 8b, Claude Haiku, and Gemini 1.5 Flash, highlighting its superior speed, efficiency, and intelligence. Before its launch, the model underwent testing in the LMSYS.org chatbot arena to ensure competitiveness.