RLM-Cascade: Response-Level Speculative Decoding for Cost-Efficient LLM API Serving

Add code
Jun 22, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: