Fable vs. Astra
- 发布
- 09/05 08:06
- 收录
- 09/05 12:00
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 正常
讨论趋势
百分比基于采集到的讨论信号，不代表新增评论数或独立参与人数。曲线仅用于同一话题在不同时段的比较。
一位开发者对OpenAI的新模型GPT-6 Astra和Fable 5.1进行了对比测试，发现Fable 5.1在处理复杂任务和维持大型上下文窗口方面，不仅成本更低，而且更加可靠。该开发者指出，使用Astra在30分钟内产生了60美元的API费用，而之前使用Fable 5的最大账单仅为8美元。目前，该开发者正在考虑是购买专业版订阅还是继续使用API来完成现有工作，因为仍需要进行8-9次调用才能获得完整的蓝图。
For those who use both top-tier models, what's your experience for both models?
Just wanted to test both frontier models "with same complex task and with same prompt". The fable 5.1 token consumption and ability to maintain extremely large context window is significantly cheaper and reliable than open Ai's new Gpt - 6 Astra model.
I've been using all the frontier models for many months In API platform and would dump and receive tokens of length 250k on average with max 3 turns(back and forth), since my actual work is occasional lengthy sessions, not everyday to justify pro plan. 100 or above dollars pro plan was not for me
And today, i was shocked to see the Open AI's Astra model's token hungry character and handling extremely long context window.
Tbf, i maxed out all the available variables for both models to test it to limit. And the token consumption is just day and night. The quality of output too isnt drastically better with Astra. the sheer no of citations it pulled is genuinely deranged.
My last biggest api bill was 8 dollars with fable 5. But now i had a 60 dollar bill for just 30 mmins of Astra visiting all the observable universe. Here are the pics
Sp, now do i get pro sub or just use API to finish the ongoing work? I still need atleast 8-9 calls to get the full blueprint that can be implemented by efficient, work horse models