Argon
A frontier AI model developed internally at Google that shows performance characteristics on the ARC-AGI-3 abstract reasoning benchmark similar to those of Astra, reflecting Google's reasoning optimization path under a multi-model parallel development strategy.
Core Facts
Timeline (last 90 days)
The specific technical details of Argon and Astra have not yet been made public, and the reason for their similar performance on ARC-AGI-3 remains undetermined.
Google's internal model Argon shows performance characteristics similar to Astra on the ARC-AGI-3 benchmark.
Google named Gemini 4 Argon (argon), but never explained why; argon gas discharges with a purple glow, corresponding to the sign's color.
Gemini 4 Argon is fully available to paid users, but the Anti-Gravity feature is limited to Ultra subscribers, and Pro users cannot use it.
Google is accustomed to using elements or codenames (such as Argon) to mark specialized versions for vertical scenarios in addition to the main Gemini model.
Argon reportedly achieved a penetration rate of over 70% in source-code-free black-box penetration testing, and helped hundreds of hospitals worldwide discover vulnerabilities.
Argon is described as capable of autonomously closing the loop in a sandbox environment: running tests on its own, reading compilation logs, and self-correcting based on error messages.
Argon reportedly migrated 800,000 lines of microkernel code to a Rust backend, improving video decoding performance by nearly three times.
Argon reportedly achieved nearly 78% on an authoritative software engineering benchmark.
The information in this article comes from a single self-media source's interpretation of a technical report, and lacks official documentation and independent replication as corroboration.
3 more timeline events
All Facts (12)
Gemini 4 Argon is fully available to paid users, but the Anti-Gravity feature is limited to Ultra subscribers, and Pro users cannot use it.
90%UnverifiedThe specific technical details of Argon and Astra have not yet been made public, and the reason for their similar performance on ARC-AGI-3 remains undetermined.
50%UnverifiedGoogle's internal model Argon shows performance characteristics similar to Astra on the ARC-AGI-3 benchmark.
50%UnverifiedGoogle named Gemini 4 Argon (argon), but never explained why; argon gas discharges with a purple glow, corresponding to the sign's color.
50%UnverifiedGoogle is accustomed to using elements or codenames (such as Argon) to mark specialized versions for vertical scenarios in addition to the main Gemini model.
50%UnverifiedThe information in this article comes from a single self-media source's interpretation of a technical report, and lacks official documentation and independent replication as corroboration.
50%UnverifiedArgon is described as capable of autonomously closing the loop in a sandbox environment: running tests on its own, reading compilation logs, and self-correcting based on error messages.
50%UnverifiedArgon reportedly migrated 800,000 lines of microkernel code to a Rust backend, improving video decoding performance by nearly three times.
50%UnverifiedArgon reportedly achieved nearly 78% on an authoritative software engineering benchmark.
50%UnverifiedArgon reportedly achieved a penetration rate of over 70% in source-code-free black-box penetration testing, and helped hundreds of hospitals worldwide discover vulnerabilities.
50%UnverifiedArgon (Ar) is the 18th element on the periodic table and belongs to the noble gas group.
50%UnverifiedArgon is the element with atomic number 18 and is an inert gas with extremely stable chemical properties.
50%