OpenAI “慧眼”识图！全新o3/o4-mini模型登场

OpenAI's "Sharp Eyes" for Image Recognition! The All-New o3/o4-mini Models Debut 🎉

AI 圈又热闹了！OpenAI 近日发布了新一代推理模型 o3 和 o4-mini，最抓人眼球的莫过于它们学会了"看图说话"。这可不是简单的图像识别，而是像人类工程师一样，能对设计图纸进行推敲、理解。OpenAI 的 CEO Sam Altman 对此赞不绝口，直呼它们"接近天才水平"。

The AI world is buzzing again! OpenAI recently released its new generation of reasoning models, o3 and o4-mini. What's most eye-catching is that they've learned to "describe images." 🤩 This isn't just simple image recognition; it's like human engineers being able to scrutinize and understand design blueprints. 🤯 OpenAI CEO Sam Altman raved about them, calling them "close to genius level." 🧠✨

三大突破性进化：不止是"聪明"，更是"多才多艺"

Three Groundbreaking Evolutions: More Than Just "Smart," But Also "Versatile" ✨

• 视觉推理：AI 也能"眼见为实"

  o3 和 o4-mini 最大的亮点，就是拥有了视觉推理能力。模型不仅能理解图像，还能基于图像进行分析推理，甚至可以像我们用手势操作手机屏幕一样，动态调整图像。

• Visual Reasoning: AI Can Also "See and Understand"

  The biggest highlight of o3 and o4-mini is their visual reasoning capability. The models can not only understand images, but also perform analysis and reasoning based on images, and even dynamically adjust images like we use gestures to operate a mobile phone screen.

• 自主Agent：AI界的"瑞士军刀"

  想象一下，AI 像一个经验丰富的项目经理，能自主判断并组合运用各种工具，解决复杂问题。据说，为了搞定一个棘手的任务，模型曾连续调用了近 600 次工具！

• Autonomous Agents: The "Swiss Army Knife" of AI

  Imagine AI as a seasoned project manager, capable of independently deciding and combining various tools to solve complex problems. Reportedly, to tackle one tricky task, the model called upon nearly 600 tools in a row! 🤯

• 性能飞跃：学霸模式全开

  o3 作为 OpenAI 目前最强的模型，在编程、数学、科学等领域都取得了显著提升。而 o4-mini 则更注重效率和性价比，在非 STEM 领域和数据科学方面表现亮眼。

• Performance Leap: Scholar Mode Activated 🚀

  o3, as OpenAI's most powerful model to date, has achieved significant improvements in fields such as programming, mathematics, and science. Meanwhile, o4-mini focuses more on efficiency and cost-effectiveness, excelling in non-STEM fields and data science. ✨

如何体验？OpenAI 的"下一步棋"

How to Try It Out? OpenAI's "Next Move" 🚀

目前，ChatGPT Plus、Pro 和 Team 用户已经可以尝鲜 o3、o4-mini 和 o4-mini-high 了。免费用户也能在编辑器中选择"Think"模式，体验 o4-mini 的魅力。开发者则可以通过 API 接入。OpenAI 预计几周后将发布 o3-pro，并提供更全面的工具支持。

Currently, ChatGPT Plus, Pro, and Team users can already experience o3, o4-mini, and o4-mini-high. Free users can also select the "Think" mode in the editor to enjoy the charm of o4-mini. Developers can access it through the API. OpenAI anticipates releasing o3-pro in a few weeks and providing more comprehensive tool support.

此外，OpenAI 似乎还在酝酿一个"社交梦"， 正在探索类似 X 的社交网络项目，将 ChatGPT 的图像生成功能与社交信息流相结合。不过，目前还处于早期阶段——未来会如何发展，让我们拭目以待。

In addition, OpenAI seems to be brewing a "social dream," exploring a social network project similar to X, combining ChatGPT's image generation capabilities with a social feed. However, it is currently in the early stages—let's wait and see how it develops in the future. 👀

🧠 收藏➕关注 每日掌握前沿科技，同步提升英语硬实力！科技英语双丰收！🎉

🧠 Collect ➕ Follow to master cutting-edge technology daily and improve your English skills simultaneously! Reap the benefits of both technology and English! 🎉