TokenFlyUsers & pricing用户与定价
The multiplier is what a user pays as a multiple of our upstream cost, on every model. 1.08 means cost + 8%. Leave it empty to use the default. Rates with a known maker's list price have a floor of 20% of list, remain capped at list and never go below cost. The multiplier range is . 倍率是用户在所有模型上支付的价格相对上游成本的倍数，1.08 即成本加 8%。留空表示使用默认倍率。已知厂商官方价的单价最低为官方价的 2 折，仍不高于官方价、不低于成本。倍率可设范围为 。
| User用户 | Joined注册时间 | Balance余额 | Spend, 30d近 30 天消费 | Requests, 30d近 30 天请求 | Keys | Region账户区域 | Foreign models国外模型开白 | Multiplier (all models)倍率（所有模型） |
|---|---|---|---|---|---|---|---|---|
| Loading…加载中… | ||||||||
ProvidersProvider 管理
Every upstream TokenFly buys from. A provider switched off is skipped: requests go to the next one in each model's order. Balances come from the provider where it offers them, refreshed every minute; health is over the last 24 hours. TokenFly 接入的所有上游。停用的 Provider 会被跳过，请求按各模型的顺序交给下一个。余额由 Provider 提供（每分钟刷新，部分 Provider 不提供），健康度为近 24 小时。
| Provider | Balance余额 | Models模型数 | Health, 24 h近 24 小时健康度 | Status状态 |
|---|---|---|---|---|
| Loading…加载中… | ||||
Models & providers模型与 Provider
Providers are listed in the order they are tried; a request moves to the next one when a provider fails. By default preferred providers come first, then the cheapest (default, which sets the price) and the rest, fallback providers last. Use ↑ ↓ to set a model's own order; it changes who serves the request, not the catalog price. Health = successful calls / all calls, retries included. Prices are upstream cost in USD per 1M tokens (input / output). Provider 按尝试顺序排列，前一个失败时自动尝试下一个。默认顺序：优先 Provider 最先，其次是最便宜的（默认，售价按它计算）和其他 Provider，兜底 Provider 最后。可用 ↑ ↓ 为单个模型设置自己的顺序；它只改变由谁出结果，不改变目录售价。健康度 = 调用成功次数 / 总调用次数（含重试）。价格为上游成本，美元 / 百万 tokens（输入 / 输出）。
| Model模型 | Providers · price · healthProvider · 价格 · 健康度 | Overall health总体健康度 |
|---|---|---|
| Loading…加载中… | ||