From 99c0b62f1e68bfe37e975fa7405f5c0794f1fafd Mon Sep 17 00:00:00 2001 From: Opher Lieber Date: Thu, 30 Apr 2026 01:12:06 -0700 Subject: [PATCH] allow requests with exactly context_len total tokens (#22546) Co-authored-by: Ethan (Yusheng) Su --- python/sglang/srt/managers/tokenizer_manager.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/python/sglang/srt/managers/tokenizer_manager.py b/python/sglang/srt/managers/tokenizer_manager.py index acb3c9fcd..e2830217c 100644 --- a/python/sglang/srt/managers/tokenizer_manager.py +++ b/python/sglang/srt/managers/tokenizer_manager.py @@ -842,7 +842,7 @@ class TokenizerManager(TokenizerControlMixin, TokenizerManagerScoreMixin): if ( self.validate_total_tokens and max_new_tokens is not None - and (max_new_tokens + input_token_num) >= _max_req_len + and (max_new_tokens + input_token_num) > _max_req_len ): if self.server_args.allow_auto_truncate: logger.warning(