deepseek-v4.1POST
Analysis: cache reads are 96–99% of agentic coding cost, cheap open weights like V4.1 Flash drive cuts
A post citing an analysis says 96–99% of agentic coding cost is cache reads, and price cuts come mainly from cheap open-weight models like DeepSeek V4.1 Flash plus competition among inference providers, with OpenAI and Anthropic cutting up to 80%.
su8 / denchu
@__su888Archived 2 days agoOriginally posted 2026.10.27
su8 / denchu @__su888Originally posted 2026.10.27
agentic codingでは費用の96-99%がキャッシュ読込で、その値下げが効く。著者はDeepSeek V4.1 Flash等の安価なオープンウェイトと多数の推論事業者の競争が値下げの主因とみる / OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why? https://minimaxir.com/2026/10/cheap-llms/
x.com/__su888/status/2114842255345487909Archived 2026.10.27 12:09