FT Reports ByteDance Pre-Training 10T Parameter Model
Posts cite Financial Times report on ByteDance pre-training a large AI model.
X users quote a Financial Times article claiming ByteDance is pre-training an AI model with up to 10 trillion parameters. The size is said to be three times larger than Moonshot’s Kimi K3 at 2.8 trillion parameters and close to Anthropic’s Mythos system. Several posts note the reported figure rose from an earlier 5 trillion parameter rumor. The linked FT piece is presented as the source for the details shared in the thread.
Combined views
1M
24 posts, first seen 19d ago
Reactions
24 postsLA 
Lisan al Gaib@scaling01ByteDance is going for a 5T model (im sorry I included google in my AGI tier list half a year go. i was blinded by the hopium, and had longer timelines before Mythos)
- likes: 1
- replies: 0
- bookmarks: 0
- reposts: 0
LA Lisan al Gaib@scaling01ByteDance is going for a 5T model (im sorry I included google in my AGI tier list half a year ago. i was blinded by the hopium, and had longer timelines before Mythos)
- likes: 306
- replies: 18
- bookmarks: 44
- reposts: 3
ZE RS Robert Scoble@Scobleizer@zephyr_z9 You are right. Alibaba and others are working on it.
- likes: 0
- replies: 0
- bookmarks: 0
- reposts: 0
JU Jukan@jukan05FT: BYTEDANCE IS IN THE PRE-TRAINING STAGE OF A MODEL WITH UP TO 10 TRILLION PARAMETERS, APPROACHING MYTHOS IN SCALE.
- likes: 1.4K
- replies: 61
- bookmarks: 216
- reposts: 123
ZE Zephyr@zephyr_z9They will probably distill it Serving 10T models doesn't make sense right now
- likes: 274
- replies: 23
- bookmarks: 25
- reposts: 6