Grok 4.5: Einstein Intelligence, McDonald's Price! #shorts

该视频讨论了当前人工智能模型的格局，并基于一个概念图表对其进行评估。该图表衡量的是智能水平（越高越好）与成本（越靠左越便宜）之间的关系。 以下是主要信息和要点： * **理想的AI模型位置：** AI模型的最佳位置被描述为图表的左上方——以最低的成本实现最高的智能水平。 * **识别出“最佳区域”：** 图表上的一个“绿色方框”代表了模型的“最佳区域”，它提供了良好到非常高的智能水平，同时保持价格合理（位于成本轴的左半部分）。 * **当前智能上限：** 该图表上观察到的绝对最佳智能水平目前约为60%。 * **位于“最佳区域”的关键模型：** 有几个模型因处于这个理想的“绿色方框”中而被突出强调，它们以“麦当劳的薪水”提供了“接近爱因斯坦水平的智能”。这些模型包括： * Grok 4.5 * Spark 1.1 * GPT 5.6 * Luna (版本未说明) * **Grok 4.5 表现突出：** Grok 4.5 被特别指出在有意义的、价格实惠的模型中，其智能表现“碾压群雄”。 * **来自SpaceX AI团队的“实力展现”：** Grok 4.5 的表现被描述为SpaceX AI团队的“一次强烈的实力展示”，表明其对行业产生了重大影响。 * **Grok 4.5 的改进：** 据报道，Grok 4.5 比以前的版本快得多，在令牌使用方面效率高得多，并且实际成本也便宜得多。 * **AI模型的未来趋势：** 总体预测是，AI模型将“以非常快的速度变得更快、更智能，并且使用成本更低。” * **消费者受益：** 这种快速改进被视为“所有消费者的一大胜利”。 * **最佳性价比（截至录制时）：** Grok 4.5 被认为是截至录制时性价比最高的选择。

This video discusses the current landscape of AI models, evaluating them based on a conceptual chart that measures intelligence (higher up) against cost (further left, cheaper). Here are the key news items and takeaways: * **Ideal AI Model Position:** The optimal place for an AI model is described as the top-left of the chart – maximum intelligence for the lowest possible cost. * **"Sweet Spot" Identified:** A "green box" on the chart represents the sweet spot for models, offering good to very high intelligence while remaining affordable (on the left half of the cost axis). * **Current Intelligence Ceiling:** The absolute best intelligence observed on this chart is currently around the 60% mark. * **Key Models in the Sweet Spot:** Several models are highlighted for being in this desirable "green box," offering "near Einstein level intelligence for a McDonald's salary." These include: * Grok 4.5 * Spark 1.1 * GPT 5.6 * Luna (version unspecified) * **Grok 4.5 as a Standout:** Grok 4.5 is specifically singled out as "crushing on intelligence" among meaningfully affordable models. * **"Huge Flex" from SpaceX AI:** Grok 4.5's performance is described as a "huge flex from the team at SpaceX AI," indicating its significant impact on the industry. * **Grok 4.5's Improvements:** Grok 4.5 is reported to be "much faster, much more efficient in terms of token usage, and effectively much cheaper" than previous iterations. * **Future Trend for AI Models:** The overarching prediction is that AI models will "continue at a very rapid rate to get faster, smarter, and cheaper to use." * **Consumer Benefit:** This rapid improvement is framed as a "huge win for all consumers." * **Best Value (as of recording):** Grok 4.5 is presented as the best value option at the time of the recording.