OpenAI launches GPT-4o Mini, a cheaper, lighter version of GPT-4o

GPT-4o Mini launch: OpenAI back in May this year launched its flagship AI model, GPT-4o or GPT-4 omni. Now, just a few months later, the company has launched another AI model dubbed as GPT-4o Mini. This new AI model is based on GPT-4o and it is the company’s most cost-efficient small model yet.

Despite being cheaper and lighter, GPT-4o Mini doesn’t compromise in terms of performance. OpenAI says that GPT-4o Mini outperforms GPT-3.5 Turbo and other smaller models on almost all benchmarks. “GPT-4o mini surpasses GPT-3.5 Turbo and other small models on academic benchmarks across both textual intelligence and multimodal reasoning, and supports the same range of languages as GPT-4o,” the company wrote in a blog post.

GPT-4o mini scores highly on textual intelligence and reasoning (82% on MMLU), math, coding, and multimodal reasoning. pic.twitter.com/zcL2bLfDqz

— OpenAI Developers (@OpenAIDevs) July 18, 2024

What this graph says is that GPT-4o Mini not only performs significantly better when it comes to tasks pertaining to reasoning and multi-modal reasoning, but it also does significantly well in terms of math and coding proficiency.

GPT-4o Mini features

— GPT-4o Mini is OpenAI’s lightest and most cost-efficient model and it is expected to increase the range of apps built using AI. The company says that GPT-4o Mini is over 60 percent cheaper than GPT-3.5 Turbo. While GPT-3.5 Turbo costs $0.70 per one million tokens, GPT-4o Mini costs just $0.24 per one million tokens.

— GPT-4o Mini is also a lot faster than GPT-3.5 Turbo. It enables apps that chain or parallelize multiple model calls, interact with customers through fast and real-time text responses.

— At the moment, GPT-4o Mini supports text and vision in the API. It will get support for text, image, video and audio inputs and outputs coming in the future.

— GPT-4o Mini surpasses GPT-3.5 Turbo and other small models on academic benchmarks across both textual intelligence and multimodal reasoning.

— GPT-4o Mini supports the same range of languages as GPT-4o.

— OpenAI says that GPT-4o mini has the same safety mitigations built-in as GPT-4o.

GPT-4o Mini pricing and availability

OpenAI says that GPT-4o mini is now available as a text and vision model in the Assistants API, Chat Completions API, and Batch API. Developers pay 15 cents per 1M input tokens and 60 cents per 1M output tokens for the same.

The company also said that users will be able to access GPT-4o Mini in ChatGPT, Free, Plus in place of GPT-3.5 starting today.