diff --git a/mteb-zh/generate_report.py b/mteb-zh/generate_report.py
index cf0d5c7..2d36d96 100644
--- a/mteb-zh/generate_report.py
+++ b/mteb-zh/generate_report.py
@@ -34,7 +34,7 @@ class ReportMainScore:
 task_mapping[TaskType.all] = {k: v for _, mapping in task_mapping for k, v in mapping.items()}
 
 
-def generate_report_csv(results_dir: Path, task_type: TaskType = TaskType.classification):
+def generate_report_csv(results_dir: Path, task_type: TaskType = TaskType.Classification):
     scores = defaultdict(list)
     output_file: Path = Path(f'mteb-zh-{task_type.value}.csv')
 
diff --git a/mteb-zh/mteb_zh/tasks.py b/mteb-zh/mteb_zh/tasks.py
index 9a0f3f1..d2d26bd 100644
--- a/mteb-zh/mteb_zh/tasks.py
+++ b/mteb-zh/mteb_zh/tasks.py
@@ -16,10 +16,10 @@
 
 
 class TaskType(str, Enum):
-    classification = 'classification'
-    reranking = 'reranking'
-    retrieval = 'retrieval'
-    all = 'all'
+    Classification = 'Classification'
+    Reranking = 'Reranking'
+    Retrieval = 'Retrieval'
+    All = 'All'
 
 
 class TNews(AbsTaskClassification):
diff --git a/mteb-zh/run_mteb_zh.py b/mteb-zh/run_mteb_zh.py
index 2d93715..090dbf2 100644
--- a/mteb-zh/run_mteb_zh.py
+++ b/mteb-zh/run_mteb_zh.py
@@ -1,7 +1,8 @@
 from pathlib import Path
+from typing import Annotated
 
 import typer
-from mteb import MTEB
+from mteb import MTEB, AbsTask
 
 from mteb_zh.models import load_model, ModelType
 from mteb_zh.tasks import (
@@ -17,47 +18,32 @@
 )
 
 
+default_tasks: list[AbsTask] = [
+    TYQSentiment(),
+    TNews(),
+    JDIphone(),
+    StockComSentiment(),
+    GubaEastmony(),
+    IFlyTek(),
+    T2RReranking(2),
+    T2RRetrieval(10000),
+]
+
+
 def main(
-    model_type: ModelType,
+    model_type: Annotated[ModelType, typer.Option()],
     model_name: str | None = None,
-    task_type: TaskType = TaskType.classification,
+    task_type: TaskType = TaskType.Classification,
     output_folder: Path = Path('results'),
 ):
     output_folder = Path(output_folder)
     model = load_model(model_type, model_name)
 
-    match task_type:
-        case TaskType.classification:
-            tasks = [
-                TYQSentiment(),
-                TNews(),
-                JDIphone(),
-                StockComSentiment(),
-                GubaEastmony(),
-                IFlyTek(),
-            ]
-        case TaskType.reranking:
-            tasks = [
-                T2RReranking(2),
-            ]
-        case TaskType.retrieval:
-            tasks = [
-                T2RRetrieval(10000),
-            ]
-        case TaskType.all:
-            tasks = [
-                TYQSentiment(),
-                TNews(),
-                JDIphone(),
-                StockComSentiment(),
-                GubaEastmony(),
-                IFlyTek(),
-                T2RReranking(2),
-                T2RRetrieval(100000),
-            ]
-        case _:
-            raise ValueError(f'Unknown task type: {task_type}')
-    print(tasks)
+    if task_type is TaskType.All:
+        tasks = default_tasks
+    else:
+        tasks = [task for task in default_tasks if task.description['type'] == task_type.value]
+
     evaluation = MTEB(tasks=tasks)
     evaluation.run(model, output_folder=str(output_folder / model_type.value))