Google unveils Gemini 4 Argon with SOTA score on DeepSWE
Google is keeping Gemini 4 Argon behind a phased rollout while it tests the model’s cyber and agent-safety limits.
The model is first going to trusted cyber defenders through Google’s Fairwind Program, with broader access planned for paid API customers and Google AI Ultra subscribers. Google says Argon scored 77.9% on DeepSWE v1.1, tied for first on CWE-bench v1, and is already being used internally for coding, research, writing, Rust migrations and data-center optimization. Introductory pricing is set at $2 per million input tokens and $10 per million output tokens, rising later to $4 and $20. Google also says it is expanding Argon’s output limit from 64K to 1 million tokens. TestingCatalog's note
The model is first going to trusted cyber defenders through Google’s Fairwind Program, with broader access planned for paid API customers and Google AI Ultra subscribers. Google says Argon scored 77.9% on DeepSWE v1.1, tied for first on CWE-bench v1, and is already being used internally for coding, research, writing, Rust migrations and data-center optimization. Introductory pricing is set at $2 per million input tokens and $10 per million output tokens, rising later to $4 and $20. Google also says it is expanding Argon’s output limit from 64K to 1 million tokens. TestingCatalog's note
score 9