2026-09-23
GPT-6 Sol and Luna: Lower API Prices, Better Model Routing
GPT-6 Sol halves GPT-5.6 Sol token rates. Luna cuts input rates in half and output by about 58%. We examine where each model fits in production work.
2026-09-23
Claude Opus 5.5: What Its Lower API Price Changes
Opus 5.5 cuts standard input and output rates by 20% and cache reads by 60%. Here is how to measure the actual saving in long-running AI work.
2026-07-17
Kimi K3 Adoption: Workflow Tests and Approval Gates
Kimi K3 narrows the open-weight capability gap, but enterprises still need workflow tests, deployment controls, and approval gates.