Comparison · Updated 2026-07-08
Genie, Cosmos and V-JEPA: three routes
The world-model label now covers very different bets. Genie, Cosmos and V-JEPA are useful because they make the split visible.
English
If every lab uses "world model" in the same way, the term becomes useless. These three projects point to three different meanings.
Genie: generated worlds you can act inside
The Genie route is closest to the intuitive picture: create an environment, let an agent or user move through it, and keep the world responsive over time. The key question is whether this interaction remains stable enough for training and evaluation, not only for short demos.
Cosmos: infrastructure for physical AI
Cosmos is better understood as a platform layer. It combines world foundation models, data processing, tokenization, guardrails and simulation workflows for robotics and autonomous vehicles. The question is not whether it behaves like a consumer app. The question is whether developers can plug it into real training, validation and synthetic-data pipelines.
V-JEPA: predictive representations
V-JEPA is less visual as a product story because it does not center on generating pretty pixels. It predicts in representation space, aiming to learn useful structure for understanding, planning and robot control. That makes it easier to underestimate and harder to evaluate through demos.
The practical takeaway: ask what the model predicts, whether it is action-conditioned, how it is evaluated, and what downstream system can use it.
中文
如果所有实验室都用同一种方式说“世界模型”，这个词就会失去解释力。Genie、Cosmos 和 V-JEPA 之所以适合放在一起，是因为它们把三种不同含义暴露出来了。
Genie：可以行动其中的生成世界
Genie 路线最接近大众直觉：生成一个环境，让智能体或用户在里面移动，并让世界随时间保持响应。 关键问题不是演示是否惊艳，而是这种交互是否足够稳定，能不能用于训练和评估。
Cosmos：物理 AI 的基础设施
Cosmos 更适合被理解为平台层。它把世界基础模型、数据处理、tokenizer、guardrails 和仿真工作流 组合起来，服务机器人和自动驾驶。问题不是它像不像一个消费级应用，而是开发者能否把它接入真实的 训练、验证和合成数据流程。
V-JEPA：预测表征
V-JEPA 的产品叙事没那么直观，因为它不是以生成漂亮像素为中心。它在表征空间里预测，目标是学习 对理解、规划和机器人控制有用的结构。这让它更容易被低估，也更难通过演示视频来评价。
实用结论：先问模型预测什么、是否受动作条件控制、如何评测，以及哪个下游系统真的能使用它。