在人工智能领域掀起的一阵风暴中，OpenAI公司宣布了一项令人振奋的更新：GPT-4o Mini，一款专为大规模部署和广泛任务优化的AI大模型。这一创新旨在通过大幅度降低价格，使得先进的人工智能技术更加普及，惠及更多企业和项目。
GPT-4o Mini模型在MMLU测试中取得了82%的高分，这不仅在聊天领域表现出色，而且在LMSYS排行榜上，其在聊天方面的表现甚至超越了GPT-4。这款轻量级模型的商用价格也极具吸引力，每百万输入token的费用仅为15美分，每百万输出token的费用为60美分，相比之前的SOTA模型，价格降低了至少一个数量级，较OpenAI此前的旗舰产品GPT-3.5 Turbo更是便宜了60%以上。
OpenAI首席执行官山姆·奥特曼对此表示：“通往智能的成本已经变得如此低廉，以至于我们可以说它‘太便宜以至于可以计量’。”这一宣言不仅揭示了AI技术的最新进展，更预示了人工智能普及化的新纪元。
GPT-4o Mini的低成本和高效性能使其成为解决复杂任务的有力工具。无论是需要同时调用多个模型以提高处理效率的应用程序，还是需要处理大量上下文信息以提供深度洞察的场景，甚至是通过实时文本响应与用户进行互动的客户支持系统，GPT-4o Mini都能发挥其独特优势。这一模型的推出，不仅标志着AI技术的又一次飞跃，更预示着未来人工智能在各个行业和领域中的广泛应用将更加广泛和深入。
OpenAI的这一创新举措，无疑将对全球的AI生态系统产生深远影响，推动人工智能技术的普及，加速其在社会各个层面的应用与创新，为人类带来更多的可能性与便利。
英语如下：
News Title: “GPT-4o Mini by OpenAI: Affordable, Powerful, and Accessible AI Model Launches”
Keywords: GPT-4o Mini, API pricing reduction, free availability
News Content: Amidst a whirlwind in the AI domain, OpenAI announced an exhilarating update: GPT-4o Mini, an AI megamodel optimized for large-scale deployment and versatile tasks. This innovation aims to democratize advanced AI technologies by significantly reducing costs, making sophisticated AI more accessible to a broader range of businesses and projects.
GPT-4o Mini, performing with an impressive 82% score in the MMLU test, excels not just in conversational tasks but also surpasses GPT-4 in the LMSYS rankings, particularly in the realm of chat. The commercial pricing for this lightweight model is highly competitive, with costs of just 15 cents for every million input tokens and 60 cents for every million output tokens. This represents a price reduction of at least an order of magnitude compared to the leading models, and a price drop of more than 60% compared to OpenAI’s flagship product, GPT-3.5 Turbo.
Sam Altman, the CEO of OpenAI, remarked, “The cost of getting to intelligence has become so low that we can say it’s ‘too cheap to meter’.” This statement not only highlights the latest advancements in AI technology but also foreshadows a new era of AI democratization.
GPT-4o Mini’s low cost and high performance make it a potent tool for tackling complex tasks. Whether it’s applications that require the concurrent use of multiple models to enhance processing efficiency, scenarios that demand the handling of vast amounts of contextual information to provide deep insights, or customer support systems that engage with users through real-time text responses, GPT-4o Mini leverages its unique advantages. The launch of this model marks another leap in AI technology, signaling a more widespread and profound integration of AI across industries and sectors.
OpenAI’s pioneering initiative is set to have a profound impact on the global AI ecosystem, driving the widespread adoption of AI technology and accelerating its innovation and application across all societal layers, thereby opening up numerous possibilities and conveniences for humanity.
【来源】https://www.jiqizhixin.com/articles/2024-07-19-6
Views: 0