Unverified50% confidenceSolutionExact time
Some of DeepSeek's model layers use sliding window attention, attending only to the most recent 128 tokens; SWA bounded replay replays only a short window at the end of the prompt, which is said to nearly eliminate total prefill work.
1
Sources
50%
Confidence
Long-term
Relevance
10/21/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe SWA (Sliding Window Attention) bounded replay feature introduced in DeepSeek models has been fully implemented by vLLM, and official data shows TTFT drops by about 30%.70% similarUnverifiedDeepMind's previous Enformer model has an effective input window of approximately 200 kb.66% similarUnverifiedDeepSeek's first-generation model consumed 389,000 bytes per context token, whereas the new model released this month consumes only 890 bytes per token, a 437-fold reduction.65% similarUnverifiedDeepSeek-Coder具备16K甚至更长的上下文窗口64% similarUnverifiedDeepSeek's official default thinking-effort tier is the Hi tier (short-question thinking cap of 4096 or 8192 tokens); the actual amount of work done by the same-named model under different configurations is not the same.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1082057API
curl https://kongchang.com/api/v1/knowledge/claims/1082057MCP
get_claim(id=1082057)