deepseek-v4.1POST

DeepSeek V4.1 Flash cost math: ~$0.017 per million input tokens at 91.5% cache hit

Argues that with a 91.5% cache hit rate the effective input cost is about $0.017 per million tokens, alongside the $0.75 per 1,000 calls Search/Fetch API and 552B MIT-licensed weights.

Wen

@xiaoshan0413Archived 12 days agoOriginally posted 2026.10.18

Wen @xiaoshan0413Originally posted 2026.10.18

Open original on X

DeepSeek 昨天发了 V4.1-Flash-1017：benchmark 全涨、输出更短、价格没变。缓存命中率 91.5%，折算下来有效输入成本约 $0.017/百万 token——前沿模型第一次把“便宜”做成基础设施。还开了 Search/Fetch API（$0.75/千次调用），552B 权重的 MIT 协议版本当天就上了 Hugging Face。榜单前十它进不去，但它赌的是另一件事：当智能便宜得像电，拼的是谁先把它铺进每个 agent 的默认配置里。来源：https://www.straitstimes.com/world/deepseek-pushes-out-updated-ai-model-it-says-is-faster-leaner

x.com/xiaoshan0413/status/2111655047750647884Archived 2026.10.18 16:44