Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
gallerdude
8 months ago
|
parent
|
context
|
favorite
| on:
OpenAI o3 and o4-mini
For coding, I like the Aider polyglot benchmark, since it covers multiple programming languages.
Gemini 2.5 Pro got 72.9%
o3 high gets 81.3%, o4-mini high gets 68.9%
croemer
8 months ago
|
next
[–]
Isn't it easy to train on the specific Exercism exercises that this benchmark uses?
vessenes
8 months ago
|
prev
|
next
[–]
where do you find those o3 high numbers?
https://aider.chat/docs/leaderboards/
currently has gemini 2.5 pro as the leader at, as you say, 72.9%.
re-thc
8 months ago
|
parent
|
next
[–]
It's in the OpenAI article post (OP) i.e. OpenAI ran Aider themselves.
vessenes
8 months ago
|
root
|
parent
|
next
[–]
Update: the leaderboard has o3 high + 4o tops of the charts now with 82.7%. This is a) amazing b) 20x more expensive than Gemini.
jumpCastle
8 months ago
|
prev
|
next
[–]
It was a good benchmark until it entered the training set.
asadm
8 months ago
|
prev
[–]
thanks
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
Gemini 2.5 Pro got 72.9%
o3 high gets 81.3%, o4-mini high gets 68.9%