Gemini 4 Argon (High), a reasoning model developed by Google and released on September 30, 2026, achieves a score of 53 on the Artificial Analysis Intelligence Index, significantly outperforming the median of 26 among comparable models. It supports text and image inputs, generating text outputs with a 1M token context window (~1,500 A4 pages). Pricing is set at $2.00 per 1M input tokens and $10.00 per 1M output tokens, aligning with the median rates, while a blended 7:2:1 cache-hit/input/output ratio yields $1.47 per 1M tokens. During Intelligence Index evaluations, the model produced 110M output tokens, exceeding the median of 82M, indicating higher verbosity. As a proprietary model, it employs chain-of-thought reasoning for complex tasks and is available via API through one provider. Its multimodal capabilities include image analysis, and it is benchmarked against models in the same price tier, with metrics covering reasoning, coding, and agentic workflows. The model’s cost per Intelligence Index task averages $1.99, reflecting its balance of performance and affordability relative to peers.