Google has announced the pricing structure for its new AI service, Argon. Developers will be charged $2 per million input tokens and $10 per million output tokens. This pricing model reflects the industry standard where generating responses requires more computational resources than processing input.
The cost for output tokens is five times higher than for input tokens. For instance, processing 100 million input tokens costs $200, while generating 100 million output tokens costs $1,000. This structure favors applications focused on data retrieval, classification, and summarization over long-form content generation.
Argon is expected to be positioned alongside the Gemini model family. As of late September 2026, Google has not disclosed specific details regarding the service's capabilities, context window, or official release date. The company has also not identified which internal division is developing the product.
