diff --git a/tests/tool_use/test_ernie45_moe_tool_parser.py b/tests/tool_use/test_ernie45_moe_tool_parser.py
index ee9da4fd6464..8fbbbba32538 100644
--- a/tests/tool_use/test_ernie45_moe_tool_parser.py
+++ b/tests/tool_use/test_ernie45_moe_tool_parser.py
@@ -15,7 +15,7 @@ from vllm.entrypoints.openai.protocol import (
 )
 from vllm.entrypoints.openai.tool_parsers.ernie45_tool_parser import Ernie45ToolParser
 from vllm.tokenizers import TokenizerLike
-from vllm.transformers_utils.detokenizer_utils import detokenize_incrementally
+from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.transformers_utils.tokenizer import get_tokenizer
 
 # Use a common model that is likely to be available
diff --git a/tests/tool_use/test_jamba_tool_parser.py b/tests/tool_use/test_jamba_tool_parser.py
index 2413b983fe87..c7ca024f3a76 100644
--- a/tests/tool_use/test_jamba_tool_parser.py
+++ b/tests/tool_use/test_jamba_tool_parser.py
@@ -11,7 +11,7 @@ from partial_json_parser.core.options import Allow
 from vllm.entrypoints.openai.protocol import DeltaMessage, FunctionCall, ToolCall
 from vllm.entrypoints.openai.tool_parsers.jamba_tool_parser import JambaToolParser
 from vllm.tokenizers import TokenizerLike
-from vllm.transformers_utils.detokenizer_utils import detokenize_incrementally
+from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.transformers_utils.tokenizer import get_tokenizer
 
 pytestmark = pytest.mark.cpu_test
diff --git a/tests/tool_use/test_qwen3coder_tool_parser.py b/tests/tool_use/test_qwen3coder_tool_parser.py
index 3cf1f4ef89f1..864bb0d0c06c 100644
--- a/tests/tool_use/test_qwen3coder_tool_parser.py
+++ b/tests/tool_use/test_qwen3coder_tool_parser.py
@@ -18,7 +18,7 @@ from vllm.entrypoints.openai.tool_parsers.qwen3coder_tool_parser import (
 )
 from vllm.entrypoints.openai.tool_parsers.qwen3xml_tool_parser import Qwen3XMLToolParser
 from vllm.tokenizers import TokenizerLike
-from vllm.transformers_utils.detokenizer_utils import detokenize_incrementally
+from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.transformers_utils.tokenizer import get_tokenizer
 
 pytestmark = pytest.mark.cpu_test
diff --git a/tests/tool_use/test_seed_oss_tool_parser.py b/tests/tool_use/test_seed_oss_tool_parser.py
index 8e1ad5e9cedc..d94df61128c9 100644
--- a/tests/tool_use/test_seed_oss_tool_parser.py
+++ b/tests/tool_use/test_seed_oss_tool_parser.py
@@ -16,7 +16,7 @@ from vllm.entrypoints.openai.protocol import (
 )
 from vllm.entrypoints.openai.tool_parsers.seed_oss_tool_parser import SeedOssToolParser
 from vllm.tokenizers import TokenizerLike
-from vllm.transformers_utils.detokenizer_utils import detokenize_incrementally
+from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.transformers_utils.tokenizer import get_tokenizer
 
 pytestmark = pytest.mark.cpu_test
diff --git a/tests/tool_use/test_xlam_tool_parser.py b/tests/tool_use/test_xlam_tool_parser.py
index a1852c368eeb..fdcdd4038131 100644
--- a/tests/tool_use/test_xlam_tool_parser.py
+++ b/tests/tool_use/test_xlam_tool_parser.py
@@ -14,7 +14,7 @@ from vllm.entrypoints.openai.protocol import (
 )
 from vllm.entrypoints.openai.tool_parsers.xlam_tool_parser import xLAMToolParser
 from vllm.tokenizers import TokenizerLike
-from vllm.transformers_utils.detokenizer_utils import detokenize_incrementally
+from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.transformers_utils.tokenizer import get_tokenizer
 
 pytestmark = pytest.mark.cpu_test
diff --git a/tests/utils_/test_argparse_utils.py b/tests/utils_/test_argparse_utils.py
index 0ea4a43d2602..2d969b8c9347 100644
--- a/tests/utils_/test_argparse_utils.py
+++ b/tests/utils_/test_argparse_utils.py
@@ -10,7 +10,7 @@ import yaml
 from transformers import AutoTokenizer
 from pydantic import ValidationError
 
-from vllm.transformers_utils.detokenizer_utils import convert_ids_list_to_tokens
+from vllm.tokenizers.detokenizer_utils import convert_ids_list_to_tokens
 
 from vllm.utils.argparse_utils import FlexibleArgumentParser
 from ..utils import flat_product
diff --git a/vllm/transformers_utils/detokenizer_utils.py b/vllm/tokenizers/detokenizer_utils.py
similarity index 100%
rename from vllm/transformers_utils/detokenizer_utils.py
rename to vllm/tokenizers/detokenizer_utils.py
diff --git a/vllm/v1/engine/detokenizer.py b/vllm/v1/engine/detokenizer.py
index c55240c40f6f..6c0acd9a9f59 100644
--- a/vllm/v1/engine/detokenizer.py
+++ b/vllm/v1/engine/detokenizer.py
@@ -9,7 +9,7 @@ from tokenizers.decoders import DecodeStream
 from transformers import PreTrainedTokenizerFast
 
 from vllm.logger import init_logger
-from vllm.transformers_utils.detokenizer_utils import (
+from vllm.tokenizers.detokenizer_utils import (
     TokenizerLike,
     convert_prompt_ids_to_tokens,
     detokenize_incrementally,
diff --git a/vllm/v1/engine/logprobs.py b/vllm/v1/engine/logprobs.py
index 1c8f808bc25b..599725b6de91 100644
--- a/vllm/v1/engine/logprobs.py
+++ b/vllm/v1/engine/logprobs.py
@@ -12,7 +12,7 @@ from vllm.logprobs import (
     create_prompt_logprobs,
     create_sample_logprobs,
 )
-from vllm.transformers_utils.detokenizer_utils import (
+from vllm.tokenizers.detokenizer_utils import (
     TokenizerLike,
     convert_ids_list_to_tokens,
 )