specific method · filed under inference & serving
Context compaction
A context-management strategy described as triggering compaction at 300K tokens, following the approach used in Claude model cards.
Also called context compaction at 300K tokens, Context-compaction strategy.
- sources
- 3
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
Documented in
Further reading
Picked by hand, not extracted: where to read more, not evidence for anything on this page.
- Compaction - Claude Platform Docs docs platform.claude.comthe strategy Kimi K3 says it adopts from the Claude model cards
Evidence
3 spans quoted from the sources, strongest treatment first.
We adopt a context-compaction strategy triggered at 300K tokens.
We adopt the context-compaction strategy used in the Claude model cards, triggered at 300K tokens.
We adopt a context-compaction strategy triggered at 300K tokens.
Filed alongside
Other methods under inference & serving :: context management.