Moonshot Reports Kimi Linear: Up to 75% Cache Cut, 6× Throughput
Moonshot AI reported Kimi Linear cut KV-cache use up to 75% and lifted throughput up to 6× at a million-token context while matching full attention on benchmarks.
Moonshot AI reported Kimi Linear cut KV-cache use up to 75% and lifted throughput up to 6× at a million-token context while matching full attention on benchmarks.