人工分析基准测试：谷歌的Gemini 4 Argon与OpenAI的GPT-6 Astra对比。
谷歌DeepMind的新模型Gemini 4 Argon在人工分析智能指数（Artificial Analysis Intelligence Index）中获得了53分，与OpenAI的GPT-6 Astra持平，使谷歌重新跻身顶级人工智能实验室之列。
Artificial Analysis Benchmarks Google's Gemini 4 Argon Against OpenAI's GPT-6 Astra Google DeepMind's new Gemini 4 Argon model achieved a score of 53 on the Artificial Analysis Intelligence Index, matching OpenAI's GPT-6 Astra and placing Google back among the top three artificial intelligence labs.
在AA-Omniscience基准测试中，Gemini 4 Argon的幻觉率为15%，远低于GPT-6 Astra的51%，但其事实准确率为50%，低于Astra的63%。
Gemini 4 Argon recorded a 15% hallucination rate on the AA-Omniscience benchmark, substantially lower than GPT-6 Astra's 51%, though its factual accuracy of 50% trailed Astra's 63%.
目前Gemini 4 Argon以50%的促销折扣售价1.99美元/次智能指数任务，标准定价后价格升至3.98美元。
At an introductory 50% promotional discount, Gemini 4 Argon costs $1.99 per Intelligence Index task, rising to $3.98 once standard pricing takes effect.
该模型具有100万个标记的上下文窗口、多模态输入以及自动化能力，在AutomationBench-AA测试中排名第一（78%）。它正在逐步向选定用户推出。
The model features a 1-million-token context window, multimodal inputs, and agentic capabilities ranking first on AutomationBench-AA at 78%. It is currently rolling out to select users.