原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
商品目录
探索公开商品信息,用清晰的条件缩小选择。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
高智商正品KIRO反代,缓存不保证,有部分逆向提示词注入-----High-IQ genuine KIRO reverse proxy. No guarantee of caching. Includes extra reverse-engineered prompt injection.
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
原站按任务参数或阶梯规则计价,请在原站选择具体规格。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
公开模型参考价;实际路由、规格和收费以原站为准
原站 USD 账本额度费率;钱包充值成本和订阅权益另计
充值¥1=$1,模型折后价即为人民币价格。;基础推理强度,其他推理强度和分段累进规则以原站为准
Kimi K3Moonshot's flagship long-horizon coding and knowledge-work model with a 1M-token context window and always-on reasoning.Moonshot|1.0M context|210 Credits input/1,050 Credits output/21 Credits cache read/210 Credits cache write/ 1M tokens
媲美ANTHROPIC原生订阅版本智商的KIRO,高缓存(ANTHROPIC协议,Fable=Opus)———————— KIRO, which boasts an IQ on a par with the native subscription version of ANTHROPIC, Xhigh Cache (ANTHROPIC protocol),Fable=Opus)
第三方渠道逆向模型,真假自测。———————— Third-party channel reverse-engineering model: self-test to determine authenticity.
MAX订阅渠道,支持第三方客户端(通过修改系统提示词绕过检测),CC test评分65-100%(为了兼容第三方客户端修改了系统提示词,影响评分系统,模型是100% CLAUDE)———————— MAX subscription channel, supports third-party clients (by modifying system prompts to bypass detection), CC test score 65–100 per cent (system prompts have been modified for compatibility with third-party clients, which affects the scoring system; the model is 100 per cent CLAUDE)
各种第三方渠道和自订阅plus:不限客户端(供货不稳定,断了自己记得切换到更高分组,不建议跑大型项目除非你有自己的记忆库) Various third-party channels and Self-Subscription Plus: with no client restrictions (supply is unstable; if your service is interrupted, remember to switch to a higher tier, It's not recommended to run large projects unless you have your own memory library.)
PRO渠道,codex和codex app使用,持续稳定————————PRO Channel, for use with Codex and the Codex app,, with consistent and stable performance
原站公开价目表展示精度;未将四舍五入为零的报价视为免费
原站当前分组公开价格;分组倍率已包含,未重复折算
Kimi K3 是 Kimi 迄今能力最强的旗舰模型,拥有 2.8 万亿参数,基于 KDA 混合线性注意力机制(Kimi Delta Attention)和注意力残差(Attention Residuals)技术构建,原生支持视觉理解,并拥有 100 万 token 上下文窗口。它是全球首个开源的 3 万亿级别模型,面向长程编程、知识工作和推理等前沿智能场景而设计。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。
原站公开的站内额度标价;充值折扣、规格附加费及最终扣费以原站为准。