Aravind Srinivas:发布两款 v1.2 决策模型
我们刚刚开源了另一组最先进的决策模型:
pplx-decider-flash-v1.2:4B 参数,每百万 token 1 美分(HF Decision Index 上排名第一的 4B 模型)
pplx-decider-v1.2:27B 参数,每百万 token 2 美分(HF Decision Index 总榜第 2——暂时如此,接下来会迅速改进)
两款模型均可通过 Perplexity Decisions API 使用,也以开放权重的形式在 @huggingface 上提供。
用得开心!
另外,按每百万 token 1 美分的价格,每月使用一万亿 token,最多花费 10,000 美元。我们会进一步大幅优化输入缓存和服务,让每百万输入 token 的成本降到远低于 1 美分。
对照原文
We just open-sourced another pair of state-of-the-art Decision Models: pplx-decider-flash-v1.2: 4B params, 1 cent per million tokens (top 4B model on HF Decision Index) pplx-decider-v1.2: 27B params, 2 cents per million tokens (#2 on HF Decision Index overall - for now, will improve rapidly) Both are available on the Perplexity Decisions API as well as in the form of open weights on @huggingface Enjoy!
Also, at 1 cent per million tokens, you get to consume trillion tokens a month for $10,000 at best. We will optimize the input caching and serving a lot more to subsidize this way beyond 1 cent/million input tokens.
Get started here: https://t.co/JK73JhGwRx