Plan event · 24 Sept 2026
All plan events · 31 providers tracked
Standard Computeprice
What moved
Standard Compute per-model rows borrowed at provider list rates; GLM-5.3-Flash’s cached read stored at $0.075 → GLM-5.3-Flash cached read $0.03 — the figure Z.AI publishes for GLM-5.3-Flash — on all four plans (Starter / Economy / Standard / Pro)
Cache-provenance re-verify of the Standard Compute rule: the provider resells at list rates and publishes none of its own, so every row’s cached read must trace to the page it borrows from. GLM-5.3-Flash’s $0.075 was the expired 2026-09-09 Z.AI launch-promo figure and the GLM-5.3-FlashX row’s cache column — corrected to the $0.03 Z.AI prints for GLM-5.3-Flash today, which llmgateway’s zai mapping and the CommandCode row both carry. The other 13 cached reads matched their named sources exactly; Gemini 3.1 Pro’s citation moved off ai.google.dev/pricing, which now redirects through to a Google sign-in, to the Google Cloud Gemini pricing page. Input, output and budget unchanged.
Verified 2026-09-24 · source