Unverified50% confidenceFactExact time
Large 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.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedModern LLMs routinely have dozens or even hundreds of layers, each containing multiple attention heads, making the complete attention data an enormous multi-dimensional tensor.74% similarUnverifiedIn the demo, the local agent's context window was set to a maximum of 1 million tokens.70% similarUnverifiedMistral Large 4 has a context window of 1 million tokens68% similarUnverifiedThe LongCat 2.5 preview version offers an ultra-long context window of up to 1 million tokens.68% similarUnverifiedThe context windows of mainstream coding assistants typically range from tens of thousands to hundreds of thousands of tokens.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1114952API
curl https://kongchang.com/api/v1/knowledge/claims/1114952MCP
get_claim(id=1114952)