Event ·Multi-source

Bytedance训练中国最大AI模型，参数达10万亿

据《金融时报》报道，Bytedance正在训练一个参数高达10万亿的AI模型。该规模是Moonshot旗下Kimi K3的三倍，后者目前是中国最大的模型。这一训练项目显示Bytedance在AI竞赛中投入巨大。

Current read

字节在搞10万亿参数的模型，比Kimi K3大三倍，这下有好戏看了。

11 sources67° AI heatLast updated 2026/8/7 12:54:17

Evidence

11 sources

Primary sourceDecoder

2026/8/7 12:54:17View source

Related source 1lmarena.ai

2026/8/5 14:48:18View source

Related source 2OpenRouter

2026/8/6 12:41:01View source

Related source 3GitHub

2026/8/6 17:30:47View source

Related source 4Fireworks AI

2026/8/6 17:34:00View source

Related source 5阿里通义 Qwen

2026/8/7 09:35:51View source

Related source 6量子位

2026/8/8 02:35:38View source

Related source 7LangChain

2026/8/6 17:00:40View source

Related source 8a16z

2026/8/6 18:15:55View source

Related source 9阮一峰的网络日志

2026/8/7 00:08:27View source

Related source 10IT之家

2026/8/7 01:39:42View 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