宝玉 (@dotey) · 10/16/2026, 05:32:37
Anthropic Launches Claude Opus 5.5 Fast Mode: 2.5x Speedup at Double API Cost with Separate Billing
Anthropic has officially launched the fast mode for Claude Opus 5.5, ending its research preview and removing the waitlist. The feature offers up to 2.5x faster response times without compromising model capabilities but requires separate billing for subscribers and doubles the API cost to $8/$40 per million tokens. It is currently available on Microsoft Foundry, with Google Cloud support coming soon.
SOURCE COVERAGEFull Original Report
Opus 5.5 快速模式取消排队：订阅用户要额外付费，API 价格是标准版两倍
Anthropic 宣布，Claude Opus 5.5 的快速模式结束内测，取消候补名单。还是同一个模型，能力不变，回答写出来的速度最多快 2.5 倍。
费用不算在订阅里。Pro、Max 等订阅用户在 Claude Code 里用 /fast 打开快速模式后，花费不占套餐额度，直接从“额外用量”扣，也就是套餐之外单独充值、按量付费的钱。套餐额度没用完也一样从额外用量扣，没开额外用量就用不了。团队版和企业版要管理员先开启。Claude Code 这部分规则这次没变。
用 API 的开发者按 token 付费：每百万输入 token 8 美元，每百万输出 token 40 美元，正好是标准速度 Opus 5.5 的两倍。现在加一个参数就能用，不用再申请。
微软的 Microsoft Foundry 已上线，Google Cloud 即将上线。
官方文档：https://t.co/4DftYenv96 Claude Code 快速模式说明：https://t.co/Xol9gUB1c0
↳ 引用 @ClaudeDevs (ClaudeDevs): Opus 5.5 fast mode in the Claude Platform is no longer in research preview, and we’ve removed the waitlist!
It’s the same Opus 5.5 but up to 2.5x faster, at 40 per million tokens.
It's also available in Microsoft Foundry (Hosted on Azure), with Google Cloud coming soon.
Fast mode in Claude Code remains unchanged, and uses extra usage credits.
See our docs here:
