OpenAI has launched GPT-4o mini, which according to them is their most cost-efficient small model. The company expects GPT-4o mini will significantly expand the range of applications built with AI by making intelligence much more affordable. It is priced at 15 cents per million input tokens and 60 cents per million output tokens, an order of magnitude more affordable than previous frontier models and more than 60% cheaper than GPT-3.5 Turbo.

With its low cost and latency, GPT-4o mini enables a wide range of tasks, such as applications that chain or parallelize multiple model calls, pass a large volume of context to the model, or interact with customers through fast, real-time text responses. 

GPT-4o mini supports text and vision in the API, with support for text, image, video and audio inputs and outputs coming in the future. The model has a context window of 128K tokens, supports up to 16K output tokens per request, and has knowledge up to October 2023. 

Superior textual intelligence 

OpenAI remarks that GPT-4o mini surpasses GPT-3.5 Turbo and other small models on academic benchmarks across textual intelligence and multimodal reasoning and supports the same range of languages as GPT-4o. It also demonstrates strong performance in function calling, enabling developers to build applications that fetch data or take actions with external systems, and improved long-context performance compared to GPT-3.5 Turbo.

The developers remark that the model is better than other small models at reasoning tasks involving both text and vision, scoring 82.0% on MMLU, a textual intelligence and reasoning benchmark, as compared to 77.9% for Gemini Flash and 73.8% for Claude Haiku. It excels in mathematical reasoning and coding tasks, outperforming previous small models on the market. The model also shows strong performance on MMMU, a multimodal reasoning eval, scoring 59.4% compared to 56.1% for Gemini Flash and 50.2% for Claude Haiku.

Built-in safety

GPT-4o mini has the same safety mitigations built-in as GPT-4o, assessed using automated and human evaluations. More than 70 external experts in social psychology and misinformation tested GPT-4o to identify potential risks. OpenAI believes that insights from these expert evaluations have helped improve the safety of both GPT-4o and GPT-4o mini. GPT-4o mini in the API is the first model to apply the company’s instruction hierarchy method, which helps to improve the model’s ability to resist jailbreaks, prompt injections, and system prompt extractions. This makes the model’s responses more reliable and safer to use in applications at scale.

OpenAI