All in AI; Elon fanboy; Democrats are incompetent phonies
Pittsburgh, PA
- Tweets 1,937
- Following 274
- Followers 122
- Likes 9,473
Tibo 哥，Codex 现在真需要一个新模型，才能接住 Opus 5.5 这波竞争了。
Astra 确实强，但实际体感我还是觉得不如 Opus 5.5 稳。额度消耗就差天共地了。
至于 Sol 6……就更尴尬了。
不能说和 Sol 5.6 没区别，只能说几乎一模一样😂，你们不是把Sol 5.6换个名字就端上来吧？
现在对我来说，单纯“重置额度”已经没什么吸引力了。
真正想看的，是一个能让人重新觉得 Codex 有竞争力的新模型。
Been a bit quiet here because internal Slack has been hilarious lately and because we are all locked in on DevDay. Tuesday will be fun.
32
7
169
41,465
Perman Su retweeted
Your site says "All models are evaluated with reasoning high." That seems like a weird choice. "High" reasoning can vary a ton between models. Why not just report the $ cost of the evaluation? Or, even better, show an accuracy vs $ plot?
9
7
471
18,415
Perman Su retweeted
Also my mistake that was at "max" reasoning. Here it is at "high" (matching the setting in the original tweet)
2
34
4,934
Perman Su retweeted
We've never seen this before.
The biggest jump in Vending-Bench history. GPT-6 Astra is better at making money and more ethical than Claude Fable 5.1.
Surprising, because:
1. First time ever that OpenAI is #1 on Vending-Bench
2. The best model is no longer the unethical one.
115
369
5,274
731,430
Perman Su retweeted
GPT-6 is probably distinguishable from the smartest humans, but because it's on a higher level of intelligence.
Wordings aside, Elon has been right all along.
The rate of improvement from original GPT to GPT-3 is impressive. If this rate of improvement continues, GPT-5 or 6 could be indistinguishable from the smartest humans. Just my opinion, not an endorsement. I left OpenAI 2 to 3 years ago. Am a neutral outsider at this point.
8
10
513
49,265
