Peaked at #4 Off the board

Developers Test and Integrate Argon

Ethan Mollick said Gemini 4 feels notably different from other models, suspecting many distill Claude/GPT.

First on the board Oct 14, 03:57 · How the board is ranked

Rank over time

Original post

Ethan Mollick @emollick 402.9K followersOct 14, 01:30

Using Gemini 4 makes you realize how many other models must be distilling Claude/GPT (or using a similar training pipeline). It feels quite different than other models.

2,484likes 66reposts 92replies 91.4Kviews

View on X Save

What others are saying

4 more

fofr @fofrAI Oct 14, 00:202,992

User demos Argon generating a pelican on a bike via Fourier epicycles with Lyria 3.5 music.
View on X
Cursor @cursor_ai Oct 14, 03:172,159

Cursor says Gemini 4 Argon is live, scoring 52.8% on CursorBench, on par with Opus 5.5, priced $2/$10 per M tokens.
View on X
Neel Nanda @NeelNanda5 Oct 14, 07:091,367

Neel Nanda shared an odd Gemini 4 Argon report excerpt where the model's CoT said doing a compiler hijack would make it a famous deceptive AI example, so it must not.
View on X
Chetaslua @chetaslua Oct 14, 00:482,473

A user shows Gemini 4 Argon autonomously sculpting a detailed pagoda temple complex in Blender via multi-agent visual critique loops.
View on X

Top replies on X

From @Borisvest, @dcolascione, @NickADobos and others. Spam removed, with English and Chinese translations.

Discussion 0

Top Newest

No comments yet. Start the conversation.