Event ·Multi-source

Gemini 3.6 Flash 浏览器代理基准 68%,超越 GPT-5.6 和 Sonnet 4.6

Google DeepMind 推出 Gemini 3.6 Flash 模型,面向浏览器代理任务。在 Browser Use 的 BU Benchmark 上,Gemini 3.6 Flash 得分 68%。该成绩高于 GPT-5.6-sol 的 67% 和 Sonnet 4.6 的 62%。Opus 4.8 以 74% 领先,但 Gemini 3.6 Flash 成本远低于 Opus。该模型以 Flash 定价提供前沿级的 Web 代理能力。

Current read

想低成本做浏览器自动化?Gemini 3.6 Flash 性价比炸裂,BU 基准 68% 比 GPT-5 还高,只比最贵的 Opus 差一点但便宜很多。

11 sources81° AI heatLast updated 2026/7/21 21:03:51

Evidence

11 sources
Primary sourceGoogle AI Developers
View source
Related source 1Browser Use
View source
Related source 4@koltregaskes
View source
Related source 6Philipp Schmid
View source
Related source 8marktechpost
View source
Related source 9OpenRouter
View 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.