Expired50% confidenceFactExact time
LLaMA 1 was trained on 1.4T tokens of data, and LLaMA 2 was trained on 2T tokens of data.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Related Claims
UnverifiedThe training data volume of current mainstream large models has reached tens of trillions, or even sixty trillion, tokens.71% similarVerifiedIn 2022, DeepMind published the Chinchilla paper, which corrected the optimal training ratio and indicated that, under a given compute budget, the number of model parameters and the number of training tokens should maintain an approximate ratio of 1:2063% similarUnverified训练过程可拆分为外循环（周/月单位，接收用户反馈）和内循环（小时/天单位，快速爬坡验证）61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/53862API
curl https://kongchang.com/api/v1/knowledge/claims/53862MCP
get_claim(id=53862)