Gemini Argon
A more capable reasoning AI model from Google, used as the comparison object for Gemini Hydrogen in general intelligence evaluations.
Timeline (last 90 days)
In a test of a front-end 3D wind power inspection system, Gemini 4 Hydrogen could complete the interaction, but its visuals were about the same as Gemini Argon and not as impressive as Claude Sonnet 5.5.
At the default high thinking level, Gemini Hydrogen surpassed the stronger Gemini Argon in a comprehensive intelligence benchmark.
An official Perplexity tweet claims that on Wide and Deep Research (WANDR) tasks, Fable 5.2 beat Gemini Argon by a clear margin.
Google's Gemini 4 Carbon shows almost no meaningful improvement over the existing Gemini Argon model, and its coding quality is even worse.
Google DeepMind released its new-generation flagship model Gemini Carbon, only two weeks after the launch of the previous-generation model Gemini Argon.
In the official internal comparison table, Carbon surpasses Argon on all 17 benchmark tests; the terminal capability test rose from 57.1% to 65.2%, and the software engineering cleanup test rose from 4.3% to 9.6%.
On the fifth edition of the Comprehensive Intelligence leaderboard, Gemini 4 Carbon scored 59 points, ranking first and overtaking Claude Opus 5.5, 9 points higher than Argon.
On the complex business and knowledge work benchmark leaderboard, Carbon scored 1609 points, more than 120 points above Argon's 1483 points, but it ranked below models such as Kimi and Grok.
Gemini Argon self-reports a code generation time of 65 to 80 seconds and an output of about 10,800 to 11,500 tokens, but did not provide raw logs.
Gemini Argon's original deliverable was an interactive web page (HTML/CSS/JavaScript) with no MP4 file; the video was recorded separately during evaluation.
3 more timeline events
All Facts (5)
In a test of a front-end 3D wind power inspection system, Gemini 4 Hydrogen could complete the interaction, but its visuals were about the same as Gemini Argon and not as impressive as Claude Sonnet 5.5.
50%UnverifiedAt the default high thinking level, Gemini Hydrogen surpassed the stronger Gemini Argon in a comprehensive intelligence benchmark.
50%UnverifiedAn official Perplexity tweet claims that on Wide and Deep Research (WANDR) tasks, Fable 5.2 beat Gemini Argon by a clear margin.
50%UnverifiedIn a video simulation test of double-pendulum chaotic motion, the 1017 version tracked the light points on the pendulum arms across all 500 frames and fitted the physical parameters, whereas Gemini's Argon measured the pendulum angles and pulled them back with a hidden spring.
50%UnverifiedFor visual reproduction of complex pages (such as the NASA homepage), the 1017 version is clearly inferior to Gemini's Argon.
50%