开源模型占 OpenRouter 60.8% 推理量，钱却被第三方赚走
这篇讲透了开源 AI 最尴尬的一点：模型大家都在用，赚钱的却是托管方。做 AI 应用选推理服务前先看看，能帮你理解为什么第三方托管这么便宜，也解释了各厂商的定价逻辑。
2025年5月，开放权重模型在 OpenRouter 平台贡献了 60.8% 的 token 推理量，开源模型在用户采用上明显领先。
问题在于收入归属：Qwen、DeepSeek、Llama 这类开放权重模型允许任何推理服务商托管同一份权重，第三方运营商凭借成本结构和定价策略往往能比模型发布方本身更便宜地提供服务，导致模型创造者难以从使用量中兑现收益。
这对开发者和企业用户是利好——可选供应商多、价格持续下探；但对模型厂商意味着商业模式压力，可能推动更多厂商转向闭源或捆绑服务变现。
Open-source AI is winning users, but who captures the money?
Open-weight models generated 60.8% of tokens on OpenRouter in May. Yet third-party inference providers can host these same models cheaper than their creators.
Read more: https://t.co/WTgMGdjrYq https://t.co/6xQ3tgQMlg
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include OpenRouter, open-source models, DeepSeek, Qwen.
Source attribution: curated from Tech in Asia (@Techinasia).
