Max Woolf lined up every OpenAI and Anthropic price change since June. For coding agents, he writes, cache reads are 96 ...

Max Woolf lined up every OpenAI and Anthropic price change since June. For coding agents, he writes, cache reads are 96 to 99% of token spend. Fable 5.1 kept Fable 5's $10 input and $50 output prices but cut cache reads from $1 to $0.25, which made it roughly 70 to 75% cheaper to use. He doesn't expect prices to rise again.#AI #LLMhttps://minimaxir.com/2026/10/cheap-llms/