APEX-Agents

Gemini 4 Argon benchmark evaluation results on computer screen

Gemini 4 Argon Leads APEX-Agents Benchmark as First Model Above 80 Percent

Gemini 4 Argon achieved first place on the APEX-Agents benchmark with an 82.2 percent Pass@1 score, outperforming Sonnet 5.5 by 6.7 points.

Mohamed Ghamdi ·