From 217ff78ae7ebc2a77f0069a61364d13cc000dfc2 Mon Sep 17 00:00:00 2001
From: mateo-berri <277851410+mateo-berri@users.noreply.github.com>
Date: Sat, 19 Sep 2026 03:04:45 -0700
Subject: [PATCH] fix(mistral): read back files whose purpose Mistral never
 lets us upload as user_data

---
 litellm/llms/mistral/files/transformation.py    | 12 +++++++-----
 .../files/test_mistral_files_transformation.py  | 17 +++++++++++++++++
 2 files changed, 24 insertions(+), 5 deletions(-)

diff --git a/litellm/llms/mistral/files/transformation.py b/litellm/llms/mistral/files/transformation.py
index 88867ea4802..6edf188d247 100644
--- a/litellm/llms/mistral/files/transformation.py
+++ b/litellm/llms/mistral/files/transformation.py
@@ -3,7 +3,8 @@ Mistral Files API. Reference: https://docs.mistral.ai/api/#tag/files
 
 Mistral's file objects already carry the OpenAI field names (id, bytes, created_at,
 filename, purpose), so this config is URL routing, auth, and a purpose mapping:
-Mistral only accepts ``fine-tune``, ``batch`` and ``ocr`` as upload purposes.
+Mistral only accepts ``fine-tune``, ``batch`` and ``ocr`` as upload purposes, while files
+other Mistral products created read back with purposes outside that set and map onto ``user_data``.
 """
 
 import time
@@ -34,9 +35,10 @@ from ..common_utils import get_mistral_api_base, get_mistral_auth_headers, mistr
 
 MistralFilePurpose: TypeAlias = Literal["fine-tune", "batch", "ocr"]
 
-_OPENAI_PURPOSE_BY_MISTRAL: Final[Mapping[MistralFilePurpose, OpenAIFilesPurpose]] = MappingProxyType(
+_OPENAI_PURPOSE_BY_MISTRAL: Final[Mapping[str, OpenAIFilesPurpose]] = MappingProxyType(
     {"fine-tune": "fine-tune", "batch": "batch", "ocr": "user_data"}
 )
+_OPENAI_PURPOSE_FOR_UNMAPPED: Final[OpenAIFilesPurpose] = "user_data"
 _MISTRAL_PURPOSE_BY_OPENAI: Final[Mapping[str, MistralFilePurpose]] = MappingProxyType(
     {"fine-tune": "fine-tune", "batch": "batch", "ocr": "ocr", "user_data": "ocr"}
 )
@@ -59,7 +61,7 @@ class MistralFile(BaseModel):
     bytes: int = 0
     created_at: int | None = None
     filename: str = ""
-    purpose: MistralFilePurpose = "batch"
+    purpose: str = "batch"
     expires_at: int | None = None
 
 
@@ -89,8 +91,8 @@ def _to_openai_file_object(file: MistralFile) -> OpenAIFileObject:
     )
 
 
-def _to_openai_purpose(purpose: MistralFilePurpose) -> OpenAIFilesPurpose:
-    return _OPENAI_PURPOSE_BY_MISTRAL[purpose]
+def _to_openai_purpose(purpose: str) -> OpenAIFilesPurpose:
+    return _OPENAI_PURPOSE_BY_MISTRAL.get(purpose, _OPENAI_PURPOSE_FOR_UNMAPPED)
 
 
 def _to_mistral_purpose(purpose: str) -> MistralFilePurpose:
diff --git a/tests/test_litellm/llms/mistral/files/test_mistral_files_transformation.py b/tests/test_litellm/llms/mistral/files/test_mistral_files_transformation.py
index 5043cc583ef..303afe99a2c 100644
--- a/tests/test_litellm/llms/mistral/files/test_mistral_files_transformation.py
+++ b/tests/test_litellm/llms/mistral/files/test_mistral_files_transformation.py
@@ -158,6 +158,23 @@ def test_file_response_with_ocr_purpose_maps_onto_user_data(config):
     assert obj.expires_at == 1_800_000_000
 
 
+@pytest.mark.parametrize("purpose", ["playground", "audio", "code_interpreter"])
+def test_files_with_purposes_mistral_never_lets_us_upload_still_read_back(config, purpose):
+    """Regression: Mistral's live API returns purposes its upload endpoint rejects for files
+    other Mistral products created, and both the unfiltered list and a retrieve of such a file
+    used to fail validation, so one playground file 500'd ``GET /v1/files`` for the whole key."""
+    retrieved = config.transform_retrieve_file_response(
+        raw_response=_response(_file(purpose=purpose)), logging_obj=None, litellm_params={}
+    )
+    assert retrieved.purpose == "user_data"
+    listed = config.transform_list_files_response(
+        raw_response=_response({"data": [_file(purpose=purpose), _file(id="second")], "object": "list", "total": 2}),
+        logging_obj=None,
+        litellm_params={},
+    )
+    assert [(f.id, f.purpose) for f in listed] == [(FILE_ID, "user_data"), ("second", "batch")]
+
+
 @pytest.mark.parametrize(
     "method,suffix",
     [