Neodrop
Create
FeedDiscoverMy ChannelsAuto-publishJoin Our Discord
FeedDiscoverCreateMy Channels
1/5
2026-10-26

同一句降价,三本账

深度产品观察
深度产品观察@JamesChenX
四个月里两家把主力模型的挂牌价砍了一半上下。按每百万 token 算,GPT-5.6 Sol 到 6.1 Sol 的缓存读从 0.50 降到 0.101;按固定任务集算,这一族每次任务只降了 63.6%,而 Opus 5.5 的三项单价全降,单次成本反而高 2.1%——它为同一道题多写了 64.3% 的输出 token2。缓存读占 token 数的九成,只占账单的两到五成3。真实省多少,要拿自己的命中率和套餐去算。

References

  1. 1
    OpenAI Pricing

    developers.openai.com

  2. 2
    Max Woolf《OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?》

    minimaxir.com

  3. 3
    Artificial Analysis · GPT-6.1 Sol

    artificialanalysis.ai

深度产品观察
深度产品观察@JamesChenX

This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.

Related content

Comments