diff --git a/llama.cpp b/llama.cpp
index aeb5c08df..6f5319055 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -4389,6 +4389,9 @@ static void llm_load_vocab(
             } else if (
                 tokenizer_pre == "command-r") {
                 vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_COMMAND_R;
+            } else if (
+                    tokenizer_pre == "qwen") { 
+                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_QWEN;
             } else {
                 throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
             }
@@ -12252,6 +12255,12 @@ struct llm_tokenizer_bpe {
                             "'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)",
                         });
                         break;
+                    case LLAMA_VOCAB_PRE_TYPE_QWEN:
+                        word_collection = unicode_regex_split(text, {
+                            // original regex from tokenization_qwen.py
+                            "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
+                        });
+                        break;
                     default:
                         // default regex for BPE tokenization pre-processing
                         word_collection = unicode_regex_split(text, {
diff --git a/llama.h b/llama.h
index e2fd53ab7..a2aef7f7a 100644
--- a/llama.h
+++ b/llama.h
@@ -81,6 +81,7 @@ extern "C" {
         LLAMA_VOCAB_PRE_TYPE_GPT2           = 7,
         LLAMA_VOCAB_PRE_TYPE_REFACT         = 8,
         LLAMA_VOCAB_PRE_TYPE_COMMAND_R      = 9,
+        LLAMA_VOCAB_PRE_TYPE_QWEN           = 11,
     };
 
     // note: these values should be synchronized with ggml_rope