vLLM main-2026-07-21 · Version scoped

Settings

Name	Description	Verified	Sources
max_num_batched_tokens	Maximum number of tokens to process in a scheduler iteration.	Aug 20	configuration/optimization.md
max_num_seqs	Maximum number of sequences to process in a scheduler iteration.	Aug 20	configuration/optimization.md
max_model_len	Maximum sequence length, including prompt and output tokens, used by vLLM configuration and scheduler validation.