Gemini 4 Argon
Google's Gemini 4 Argon large language model, available in the Gemini app and coding tools.
HN commenters like it for thorough, goal-focused work on hard technical problems,12 especially security and cryptanalysis research.345 The most common complaint is that it uses up usage quotas after only a few exchanges,67 and it suits long-running tasks better than quick interactive chat.89
Why people recommend it
Security research · 3
Commenters credit it with strong results in vulnerability research and cryptanalysis.
“Gemini 4 Argon made an invaluable contribution to finding the vulnerability, exploit development, attack detection, and validation of the patch.”
“It feels like it gets things done better than Opus 5.5.”
Show 1 more
Thorough, goal-focused · 2
It is described as extremely thorough, sticking to goals with less prompting and catching work other models faked.
“It feels like it gets things done better than Opus 5.5.”
“Just switched, and it seems to be much better.”
Hard research problems · 2
Users have set it on difficult open-ended problems like SHA-256 preimages and the Voynich manuscript.
Daily driver · 1
One user made it their default for tasks needing moderate smarts or lots of prose.
“it's become my daily driver for anything requiring moderate smarts or lots of prose”
Caveats and dissent
Usage limits · 2
A single question or brief conversation can exhaust a large share of the usage quota.
“When I ask the same question to Argon it is 50% of my weekly usage limit gone.”
“One brief conversation, with three exchanges, and my quota was exhausted”
Slow · 2
It is described as a little slow and better for long-running tasks than interactive chat.
“It feels like it gets things done better than Opus 5.5.”
“The model (at least on the settings I used) was not for interactive chatting. It was for longer running tasks.”
Limited availability · 2
Early users hit 404 errors during a gradual rollout, and one found it only in the chat interface.
“tried argon 4, keeps failing with 404”
“One brief conversation, with three exchanges, and my quota was exhausted”
Underwhelming intelligence · 1
One commenter found it no smarter than other frontier models.
“felt very much the same as any other Frontier, just without the intelligence”
Often compared with
Who recommends it in their own words
From the makers not counted in the ranking
“I prefer Argon over Opus 5.5 because of the clear thinking and directness.”
“I have been using this model for a bit. I would say it is a good model”
Threads where it came up
- Gemini 4 Carbon 7 mentions
- Gemini 4 Argon 5 mentions
- Gemini 4 Argon 2 mentions
- Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5 1 mention
- PHP2Shell: AI uncovers critical RCE in PHP 1 mention
- Muse Wonder 1 mention
Data: gemini-4-argon.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 15
“I have been using it on the $20 and still have 25% left”
“When I ask the same question to Argon it is 50% of my weekly usage limit gone.”
“It feels like it gets things done better than Opus 5.5.”
“One brief conversation, with three exchanges, and my quota was exhausted”
“it's become my daily driver for anything requiring moderate smarts or lots of prose”
“I used Gemini 4 Argon (High) from the Android app with the original image”
“felt very much the same as any other Frontier, just without the intelligence”
“it's possible that Argon is well-suited for the work I do but not for other tasks”
“Gemini 4 Argon made an invaluable contribution to finding the vulnerability, exploit development, attack detection, and validation of the patch.”
“Impressive!”
“Just switched, and it seems to be much better.”
“tried argon 4, keeps failing with 404”
“It's a great model.”
“Gemini 4 Argon release a few days ago was frontier a few days ago.”
“The model (at least on the settings I used) was not for interactive chatting. It was for longer running tasks.”