
刚刚,全球AI百强榜发布!ChatGPT稳坐第一,DeepSeek第三,前50有22个来自中国 (opens in a new tab)
a16z最新发布「全球Top100消费级GenAI应用榜单」,AI竞争格局逐渐稳定,中国力量全面崛起,DeepSeek、豆包、夸克等多款产品跻身前十。ChatGPT依旧领跑,谷歌Gemini紧随其后,Grok高速逆袭。整体来看,全球AI正进入多极化竞争的新阶段。
Total 483 articles

a16z最新发布「全球Top100消费级GenAI应用榜单」,AI竞争格局逐渐稳定,中国力量全面崛起,DeepSeek、豆包、夸克等多款产品跻身前十。ChatGPT依旧领跑,谷歌Gemini紧随其后,Grok高速逆袭。整体来看,全球AI正进入多极化竞争的新阶段。

爆火的神秘图像编辑模型nano-banana,终于脱掉了“香蕉皮”! 就在今天,谷歌官方认领,并表明这个模型其实是Gemini 2.5 Flash Image。

昨晚,神秘且强大的图像生成与编辑模型 nano banana 终于正式显露真身。没有意外,它果然来自谷歌,并且也获得了一个正式但无趣的名字:gemini-2.5-flash-image-preview。

Philipp Schmid describes an agent using Bash and a media viewer. Across over 350 tasks, harness scores were close while tokens and costs differed. Each task ran once, so score differences still require repeated experiments.

The Accio-led team combines visually dependent tasks, process credit assignment and alternating training. Neos-35B averages 64.7, still 2.4 percentage points behind the comparison Gemini workflow.

The API handles yes/no, multiple-choice, and scoring tasks with multimodal input and shared context. Google reports test latency and pricing; real-world calibration and costs depend on application conditions.

Google describes the new VLA model’s dexterity, factory proof of concept and safety tests, and shows how a general-purpose model helps build simulations, check data and analyze failures.

这套方案面向医疗服务方和保险方,通过连接器整理病历、预授权与福利材料,并保留人工审核。部分连接器仍在有限测试阶段。

Google DeepMind 展示了由 Gemini 4 Carbon 驱动的形式化验证案例。团队报告四周内生成约 650 万行 Lean 证明,所验证的是明确规格下的部分安全性质。

Asking the World 让模型从视频重建可运行场景,通过查询、模拟和干预获取证据。以 Gemini-3-Flash 为基础,在 CLEVRER 和 ContPhy 上分别达到 80.82% 和 70.56%。