OpenAI has unveiled its latest artificial intelligence model, GPT-4o mini, positioning it as the most cost-efficient small model in the market. This new offering is set to revolutionize the AI landscape by making advanced intelligence more accessible and affordable for a wide range of applications.
GPT-4o mini boasts impressive capabilities, scoring 82% on the MMLU (Massive Multitask Language Understanding) benchmark and currently outperforming GPT-41 on chat preferences in the LMSYS leaderboard. The model is priced at a fraction of its predecessors, costing just 15 cents per million input tokens and 60 cents per million output tokens. This pricing structure makes it an order of magnitude more affordable than previous frontier models and more than 60% cheaper than GPT-3.5 Turbo.
The new model’s low cost and latency enable a broad spectrum of applications, including those that require multiple model calls, large context volumes, or real-time text responses. Currently, GPT-4o mini supports text and vision in its API, with plans to expand to audio, video, and image inputs and outputs in the future.
Key features of GPT-4o mini include:
- A context window of 128K tokens
- Support for up to 16K output tokens per request
- Knowledge cutoff in October 2023
- Improved handling of non-English text
In benchmark tests, GPT-4o mini has demonstrated superior performance compared to other small models:
- Scored 82.0% on MMLU, outperforming Gemini Flash (77.9%) and Claude Haiku (73.8%)
- Achieved 87.0% on MGSM for math reasoning, compared to 75.5% for Gemini Flash and 71.7% for Claude Haiku
- Scored 87.2% on HumanEval for coding performance, surpassing Gemini Flash (71.5%) and Claude Haiku (75.9%)
- Demonstrated strong multimodal reasoning capabilities with a 59.4% score on MMMU
OpenAI has emphasized that GPT-4o mini incorporates the same safety mitigations as GPT-4o, with extensive testing and evaluation by both internal and external experts.
The model is now available through OpenAI’s Assistants API, Chat Completions API, and Batch API. Additionally, ChatGPT users across Free, Plus, and Team tiers will have immediate access to GPT-4o mini, replacing GPT-3.5. Enterprise users will gain access starting next week.
With its combination of advanced capabilities and affordable pricing, GPT-4o mini is poised to significantly expand the adoption of AI across various industries and applications, furthering OpenAI’s mission to make artificial intelligence accessible to all.
