星颖 发表于 2025-12-18 06:12:03

看来Gemini 3 Flash 主打的是高效低成本

如题
https://cdn3.ldstatic.com/optimized/4X/0/0/5/00579b2c9f58d54093dee0d4c8fe67d65b5fcc2f_2_690x420.png
从 ARC-AGI-1 Leaderboard上来看
Gemini 3 Flash 跟GPT 5.2算是第一梯队
但是成本省了对手一大截
一个任务只有0.174刀而已
https://cdn3.ldstatic.com/original/4X/d/9/0/d9032266e05caf96b7b83042ed46cba37c543099.png
ARC Prize (https://arcprize.org/leaderboard)
https://cdn3.ldstatic.com/optimized/3X/d/8/d8b1cd1fe3cf23f79c687bb9ca4184d442d1dd68_2_690x362.jpeg
ARC Prize - Leaderboard (https://arcprize.org/leaderboard)
The ARC-AGI Leaderboard.
Official benchmark
https://cdn3.ldstatic.com/optimized/4X/5/f/5/5f5329f1609cfa1fa44b3dd2f055aea299eae513_2_429x500.png
https://cdn3.ldstatic.com/original/4X/6/8/6/68621d6d18da314a7f19275b453656904d1d4854.png
Google – 17 Dec 25 (https://blog.google/products/gemini/gemini-3-flash/)
https://cdn3.ldstatic.com/optimized/4X/4/7/2/47221f789a4c553dafd6406d78ce560ed168864b_2_690x387.png
Gemini 3 Flash: frontier intelligence built for speed (https://blog.google/products/gemini/gemini-3-flash/)
Gemini 3 Flash offers frontier intelligence built for speed at a fraction of the cost.
页: [1]
查看完整版本: 看来Gemini 3 Flash 主打的是高效低成本