OpenAI, led by Sam Altman, has announced a new, cost-effective, small-scale AI model called GPT-4o mini, which is 60% cheaper than GPT-3.5 Turbo. GPT 4o-mini is priced at 15 cents per million input tokens and 60 cents per million output tokens.
GPT 4o-mini Features
The Microsoft-backed company said in a press release that the new GPT-4o mini can outperform GPT-4 in a chat setting on the LMSYS leaderboard, with the model scoring 82% in Massive Multitask Language Understanding (MMLU), according to OpenAI.
OpenAI's GPT-4o mini enables a wide range of tasks, including applications that chain or parallelize multiple model calls (e.g., calling multiple APIs), pass large amounts of context to models (e.g., complete code bases or conversation histories), and interact with customers through fast, real-time text responses (e.g., customer support chatbots).
Currently, GPT-4o mini supports text and vision in its API, and the company plans to add support for text, image, video, and audio input and output in the future.
According to OpenAI, GPT 4o-mini achieved a score of 87.0% in mathematical reasoning, compared to 75.5% for Gemini Flash and 71.7% for Claude Haiku. Similarly, in HumanEval, which measures coding performance, GPT-4o mini achieved a score of 87.2% compared to 71.5% for Gemini Flash and 75.9% for Claude Haiku.
Availability and Price
GPT-4o mini is available as a text and vision model in the Assistant, Chat Completion, and Batch APIs. Developers pay 15 cents per million input tokens and 60 cents per million output tokens (roughly the equivalent of 2,500 pages of a standard book).
ChatGPT's Free, Plus, and Team users will have access to GPT-4o mini instead of GPT-3.5 starting today, with Enterprise users gaining access next week.
Safety measures for GPT-4o mini
According to OpenAI, GPT-4o mini incorporates the same safety measures as GPT-4o. More than 70 external experts in fields such as social psychology and misinformation tested GPT-4o and identified potential risks. The company says that insights gained from these expert evaluations helped improve the safety of both GPT-4o and GPT-4o mini.
GPT 4o-mini Features
The Microsoft-backed company said in a press release that the new GPT-4o mini can outperform GPT-4 in a chat setting on the LMSYS leaderboard, with the model scoring 82% in Massive Multitask Language Understanding (MMLU), according to OpenAI.
OpenAI's GPT-4o mini enables a wide range of tasks, including applications that chain or parallelize multiple model calls (e.g., calling multiple APIs), pass large amounts of context to models (e.g., complete code bases or conversation histories), and interact with customers through fast, real-time text responses (e.g., customer support chatbots).
Currently, GPT-4o mini supports text and vision in its API, and the company plans to add support for text, image, video, and audio input and output in the future.
According to OpenAI, GPT 4o-mini achieved a score of 87.0% in mathematical reasoning, compared to 75.5% for Gemini Flash and 71.7% for Claude Haiku. Similarly, in HumanEval, which measures coding performance, GPT-4o mini achieved a score of 87.2% compared to 71.5% for Gemini Flash and 75.9% for Claude Haiku.
Availability and Price
GPT-4o mini is available as a text and vision model in the Assistant, Chat Completion, and Batch APIs. Developers pay 15 cents per million input tokens and 60 cents per million output tokens (roughly the equivalent of 2,500 pages of a standard book).
ChatGPT's Free, Plus, and Team users will have access to GPT-4o mini instead of GPT-3.5 starting today, with Enterprise users gaining access next week.
Safety measures for GPT-4o mini
According to OpenAI, GPT-4o mini incorporates the same safety measures as GPT-4o. More than 70 external experts in fields such as social psychology and misinformation tested GPT-4o and identified potential risks. The company says that insights gained from these expert evaluations helped improve the safety of both GPT-4o and GPT-4o mini.
