Grok 3 mini thinking is competitive with recently released models like Gemini 2.5 flash thinking and o4 mini high in same class but way cheaper

发布时间: 2025-04-19 16:30:36

1分

数据加载中

关注推特

收听电报

2

1

0

Grok 3 mini thinking is competitive with recently released models like Gemini 2.5 flash thinking and o4 mini high in same class but way cheaper
btc
( twitter.com )

昨天由 Tech Dev Notes 提交

点击图片查看原图

Markdown支持

评论加载中...

您可能感兴趣的：更多

1

2

1

1

Andrej 对 Grok 3 的评价还蛮高的：Grok 3 + Thinking 给我的感觉是：它大概和 OpenAI 最强的模型（比如 o1-pro，月费 200 美元）相当，比 DeepSeek-R1 和 Gemini 2.0 Flash Thinking 略好一些。
***
Andrej Karpathy
我今天早些时候拿到了 Grok 3
IT技术
( twitter.com)

2个月前 • 宝玉 • -- 点击 0 评论

2

2

1

1

🚨 GROK 3 MINI: THE BUDGET GENIUS IS HERE
Grok 3 Mini just ranked above DeepSeek and QwQ—and nearly matched Claude 3.7 in global performance, all while costing a fraction to run.
At $0.30 per million input tokens and $0.50 per million output, it's easily one of the cheapest
btc
( twitter.com)

4天前 • Mario Nawfal • -- 点击 0 评论

3

2

1

1

"Is 112345679 a prime number?" My friend asked me to test this on Grok 3 reasoning this morning. Unsurprisingly, Grok nailed it with sharp reasoning -- not a prime number as it’s 89 × 1262311! Btw, neither o1 nor o3-mini-high can answer it correctly.
🔗 Links below
- Grok-3
btc
( twitter.com)

1个月前 • Bill Yuchen Lin • -- 点击 0 评论

4

2

1

1

Grok-3-mini-fast really is chasing that efficient frontier. There are benchmarks and then there are results purely measured in utility.
btc
( twitter.com)

10天前 • Adam Ghetti • -- 点击 0 评论

5

2

1

1

根据 LiveBench 的评分表格，xAI 的 “grok-3-thinking” 总体平均得分 71.57%，在编程任务上得分 67.38%，低于 OpenAI 的 o3-mini-2025-01-31-high（82.74%）和 o1-2024-12-17-high（69.69%）。
LiveBench 是一个由 xAI 等机构开发的无污染、客观的 LLM
IT技术
( twitter.com)

1个月前 • 宝玉 • -- 点击 0 评论

6

2

1

1

Btw, the chain of thought in the "thinking" mode for Grok 3 is completely open. No summarizers or obfuscation. This is really important and the reasoning process is often fascinating!
btc
( twitter.com)

1个月前 • Keiran Paster • -- 点击 0 评论

7

2

1

1

Grok 3 introduced a very compelling Thinking UX. Super elegant approach for showing reasoning in real-time, and gives the user a ton of confidence in what's happening. Give the designer a raise.
btc
( twitter.com)

1个月前 • Aaron Levie • -- 点击 • 下载视频 0 评论

00:00:25

8

2

1

1

In case anyone still doesn't see the insane speed that models are getting smarter and cheaper:
Yesterday, Google released Gemini 2.5 Flash, a very efficient reasoning model.
Today, Grok 3 mini is stronger on most benchmarks for 7x cheaper!

btc
( X.com)

昨天 • Keiran Paster • -- 点击 0 评论

9

2

1

1

o3-mini-high gives up and thinks issues are unsolvable. It just told me to open an RFC lol. Grok 3 will not give up, and will actually figure out a solution or workaround. It is higher agency in this respect
btc
( twitter.com)

1个月前 • Sauers • -- 点击 0 评论

0.05903 Second , Gzip Enable.本网所有言论均来自网络，不代表本网站立场。联系方式: [email protected]

©2012.11.21 bad.news All rights reserved. 社区自动运营第 -- 年零 -- 天
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

关注推特