Skip to main content
Shipped
No vapor
The AI that actually shipped
Front
Decision models
New
Rising
Vapor log
Roundup
Submit
Sign in
Home
Wiki
KV-cache compression
Copy Markdown
View as Markdown
KV-cache compression
KV-cache compression is listed here as a concept.
2026-09-07
2026-10-19
1
Items per week for 7 weeks, 2026-09-07 to 2026-10-19; peak 1; latest week 0. Bars: items per week.
Mentioned with
Maker
DeepSeek-AI
1
Model
DeepSeek-V4.1-Flash
1
Platform
Hugging Face
1
Concept
KV cache
1
Licence
MIT
1
Task
agentic workloads
1
Task
long-context inference
1
Related coverage
DeepSeek releases V4.1-Flash, a multimodal MoE with a much smaller KV cache, under MIT
2026-09-10