Event ·Multi-source

Gemini 3.6 Flash发布，浏览器智能体任务性价比最佳

Google DeepMind发布Gemini 3.6 Flash，在BU Benchmark上达到68%准确率。该成绩超过GPT-5.6-sol（67%）和Sonnet 4.6（62%）。它仅次于Opus 4.8（74%），但成本仅为后者的零头。同时发布的还有低成本的Gemini 3.5 Flash-Lite和网络安全模型Gemini 3.5 Flash Cyber。

Current read

Google DeepMind新模型Gemini 3.6 Flash在浏览器自动化上性能接近Opus 4.8但价格便宜很多，还顺带发布了安全模型。

11 sources85° AI heatLast updated 2026/7/21 16:20:41

Evidence

11 sources

Primary sourceBrowser Use

2026/7/21 16:20:41View source

Related source 1Google AI Developers

2026/7/21 21:03:51View source

Related source 2Jeff Dean

2026/7/21 15:37:20View source

Related source 3Poe

2026/7/22 15:32:20View source

Related source 4marktechpost

2026/7/21 17:45:08View source

Related source 5Philipp Schmid

2026/7/21 15:31:47View source

Related source 6小互

2026/7/21 15:48:44View source

Related source 7宝玉

2026/7/21 17:25:31View source

Related source 8Geek

2026/7/21 14:49:28View source

Related source 9OpenRouter

2026/7/21 21:54:19View source

Related source 10@koltregaskes

2026/7/22 07:31:52View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story