OpenCode 24h API usage: DeepSeek V4.1 Flash leads with 8.5T tokens
DeepSeek V4.1 Flash 1017 tops OpenCode's 24-hour API call chart at 8.5T tokens versus 503B for Meta's muse-spark, while new Kimi K3.1 Pro misses the top 18.
🐳 The Truth About Token Usage Over the Past 24 Hours: DeepSeek V4.1 Flash 1017 firmly holds the top spot, while Kimi K3.1 Pro has completely vanished! Top 5 models by real-time API calls (@opencode): 1️⃣ DeepSeek V4.1 Flash 1017 (DeepSeek) — 8.5T (crushing the entire field) 2️⃣ muse-spark-1.3-contributor (Meta) — 503B 3️⃣ glm-5.3-flash (Zhipu) — 479B 4️⃣ beam-1.0-501b-a23b (Reflection) — 399B 5️⃣ mimo-v2.6-flash (Xiaomi) — 294B The newly released Kimi K3.1 Pro has completely disappeared from the chart, failing to even make the top 18. Meanwhile, DeepSeek V4.1 Flash 1017 continues to stand tall at the top. High cost-effectiveness + open-source weights + actual usability — developers are voting with their hard-earned money and real usage volume. This isn't marketing Moonshot. This is the real people's vote! P.S. The app, Zhipu @Zai_org, make an actual app!
This post quotes another post: x.com/lindagreyhill/status/2114361229808447720