[AI Payment Infrastructure Kite Launches Adaptive Inference Infrastructure Marathon for Long-Running AI Agents] BlockBeats News, July 22: The AI payment infrastructure Kite ecosystem has launched Marathon, an adaptive inference infrastructure designed for long-running AI Agents. Users can select a completion window for each request (four tiers: now / soon / later / anytime), with longer wait times resulting in lower prices; the deepest tier can save up to approximately 65% compared to real-time pricing (discount is an estimated value and fluctuates with real-time capacity). The service API is compatible with OpenAI and offers one-line command installation plugins for Claude Code and Codex. The initial release supports five leading open-weight models: Kimi K3, GLM 5.2, DeepSeek V4 Pro, Qwen3.6-35B-A3B, and Nemotron 3 Ultra, all under a unified pricing structure. For more details, visit the official website marathon.build. It is reported that Kite previously completed a $33 million funding round led by PayPal Ventures and General Catalyst. The launch of Marathon marks a critical step in extending from its role as an Agent and payment network to becoming the foundational computing infrastructure for the Agent economy, aiming to establish a complete closed-loop system for Agents to "autonomously invoke and autonomously pay." [Original Link]