Google ships Gemini 4 Argon to developers at an introductory $2/$10 per million tokens
The new model is available now through the API, with pricing set to double after November 30. Google's headline claims, an 87% embodied-reasoning score and a 0.7% prompt-injection success rate, are its own benchmarks.
Google released Gemini 4 Argon on October 13. Per the announcement by Koray Kavukcuoglu, SVP of Google DeepMind and Google's Chief AI Architect, the model "is now globally available today for developers and enterprises via the API," and also reaches Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search.
A launch price with an expiry date
API access starts at "$2 per million input tokens and $10 per million output tokens." That is an introductory rate: after November 30, 2026, "the price of $4 per 1M input tokens and $20 per 1M output tokens will apply." Teams that build on Argon this quarter should budget for the doubled rate, not the launch one.
The claims, and who measured them
Google says that, "relying on pure reasoning, Argon outperforms all other frontier models, with an 87% score" on the ERQA embodied-reasoning benchmark. On Gray Swan's indirect prompt-injection test, it reports "cutting attack success across 15 adaptive attempts to 0.7%." It also describes an internal robotics trial in which Argon watched raw video of a person playing cornhole, built the simulation code, and the robot "quickly sank seven consecutive shots." All three results are Google's own reporting.
Why it matters for AI spend
The price is the part that is certain: whatever Argon is worth on your workloads, it will cost twice as much per token from December. A launch discount is the cheapest time to run your own evaluation against the model you already pay for, before a migration bakes the higher rate into next year's budget.