Unverified50% confidenceFactExact time
A 10-second video at 24 frames per second with 256 tokens per frame produces more than 60,000 tokens, and the attention matrix reaches billions of elements in scale.
1
Sources
50%
Confidence
Long-term
Relevance
11/3/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe dataset consists of three core tables: the Creators table with approximately 4.5 billion rows, the Videos table with approximately 5.6 billion rows, and the Sounds table with approximately 633 million rows.71% similarUnverifiedLarge 4's context window reaches one million tokens and uses a sparse attention mechanism that attends to at most 2048 of the most relevant positions each time.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1129715API
curl https://kongchang.com/api/v1/knowledge/claims/1129715MCP
get_claim(id=1129715)