精选理由
Codex一周300元,国内大模型一天7800,差距太夸张了!重度用户选对定价模式比模型性能更关键。
重度AI编程用户使用Codex每周仅需300元订阅费,而使用国内模型如GLM-5.2每天花费高达7800元。实际开发工作流中每日token消耗达数十亿级别,按token计费的国内模型总成本远超订阅捆绑定价的Codex。订阅制降低高频使用者的边际成本,而按token计费模式导致重度用户支出爆炸性增长。
原文 · pandaily
Domestic Chinese LLMs Too Expensive for Heavy Users: Per-Token Cost Advantage Vanishes When Real Development Workflows Reach Tens of Billions of Tokens Daily
Chinese heavy AI coding users spend 300 yuan/week on Codex vs 7800 yuan/day on domestic models like GLM-5.2, as subscription bundling rather than per-token pricing determines real-world cost.