Unverified50% confidenceFactExact time
OpenAI动用Codex等模型分析了超过70亿条日志来调查该事件，耗费数百万GPU小时
1
Sources
50%
Confidence
Long-term
Relevance
8/15/2026
First Seen
Sources
Related Claims
UnverifiedA content creator consumed approximately 660 million tokens after intensively using Codex to complete multiple projects.68% similarUnverifiedCodex's Computer Use feature is extremely token-intensive, consuming nearly 80,000 context tokens for a simple task64% similarUnverifiedThe build process involved 15,314 model calls, more than 339 million input tokens, and a total of more than 2.6 billion tokens (including cached reads, output, and thinking tokens).64% similarVerified2025年初OpenAI Codex的周活用户为60万，不到4个月飙升到400万，翻了将近7倍64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/753137API
curl https://kongchang.com/api/v1/knowledge/claims/753137MCP
get_claim(id=753137)