OpenAI has introduced GPT-4o mini, a compact model aimed at making AI more accessible and affordable for developers. This new member of the GPT family offers enhanced performance at a significantly lower cost compared to earlier models, including the standard GPT-4o.
Despite its smaller size, GPT-4o mini delivers impressive capabilities, achieving an 82% score on the MMLU benchmark and outperforming GPT-4 in chat preferences on the LMSYS leaderboard. The model is priced at just 15 cents per million input tokens and 60 cents per million output tokens, making it much more economical than its predecessors.
Key Features:
- Low Cost and Latency
- 128K Token Context Window
- Up to 16K Output Tokens per Request
- Knowledge Cutoff: October 2023
- Enhanced Tokenizer for efficient handling of non-English text
- API Support for Text and Vision (with plans to expand to video and audio)
GPT-4o mini excels against other small models in several benchmarks:
- MMLU (Textual Intelligence): 82.0%
- MGSM (Math Reasoning): 87.0%
- HumanEval (Coding Performance): 87.2%
- MMMU (Multimodal Reasoning): 59.4%
These results highlight GPT-4o mini’s superiority in reasoning tasks, mathematics, coding, and multimodal understanding compared to competitors like Gemini Flash and Claude Haiku.
Developers can utilize GPT-4o mini for various applications, including:
- Chaining or parallelizing multiple model calls
- Handling large volumes of context (e.g., complete codebases or conversation histories)
- Creating real-time text response systems (e.g., customer support chatbots)
OpenAI has emphasized safety in the development of GPT-4o mini by incorporating pre-training content filtering, post-training alignment techniques like RLHF, and an innovative “instruction hierarchy” approach to prevent jailbreaks and prompt injections.
GPT-4o mini is currently available through the Assistants API, Chat Completions API, and Batch API. Developers can expect costs of 15 cents per million input tokens and 60 cents per million output tokens, with fine-tuning capabilities set to be introduced in the coming days.
OpenAI envisions a future where AI models are seamlessly integrated into every application and website. “GPT-4o mini is paving the way for developers to build and scale powerful AI applications more efficiently and affordably,” they state.
As AI technology continues to advance, GPT-4o mini represents a significant step toward making sophisticated language models more accessible to developers of all backgrounds. With its impressive performance and affordability, this new model is poised to usher in a new era of AI-driven applications and services while anticipation builds for GPT-5.