Claude Opus 5: The Rise of Level 5 Long-Horizon AI

The latest frontier AI models, including GPT-5.6, Kimi K3, and Claude Opus 5, represent a new, higher level of AI. These models are now considered "Long-horizon multi-agent AI." What's interesting is how AI has progressed through distinct levels and cycles. We've seen familiar "AI freakouts," like concerns about AI taking jobs or fears that "AI will kill us all" when powerful new models emerge. For example, a model called Mythos 5 caused such a stir that another model, Fable 5, was briefly pulled from the market. There are also "Chinese AI freakouts" whenever China releases a new frontier model, like Kimi K3, leading to discussions about an "AI race." The evolution of AI systems since late 2023 can be described in five capability levels. Level 1 was conversational AI that answered questions, like GPT-4. Level 2 brought multimodal AI that understood different types of information, seen with GPT-4o and Gemini-1.5. Level 3 introduced reasoning AI, which could solve difficult problems. Then came Level 4, agentic AI, which completes complex tasks autonomously, like Claude 4.x Opus and GPT-5. The bottom line: AI capabilities are rapidly advancing, requiring us to increase our ambition to leverage this new technology.