Merge remote-tracking branch 'origin/main'

Author: yanbasic

src/emd/models/llms/qwen.py

@@ -846,6 +846,66 @@
     )
 )
 
+Model.register(
+    dict(
+        model_id = "Qwen3-30B-A3B-Instruct-2507",
+        supported_engines=[vllm_qwen3_engin091],
+        supported_instances=[
+            g5d12xlarge_instance,
+            g5d24xlarge_instance,
+            g5d48xlarge_instance,
+            local_instance
+        ],
+        supported_services=[
+            sagemaker_service,
+            sagemaker_async_service,
+            ecs_service,
+            local_service
+        ],
+        supported_frameworks=[
+            fastapi_framework
+        ],
+        allow_china_region=True,
+        huggingface_model_id="Qwen/Qwen3-30B-A3B-Instruct-2507",
+        modelscope_model_id="Qwen/Qwen3-30B-A3B-Instruct-2507",
+        require_huggingface_token=False,
+        application_scenario="Agent, tool use, translation, summary, instruction following, reasoning",
+        description="Qwen3-30B-A3B-Instruct-2507 is an updated instruction-tuned version featuring significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage, with enhanced 256K long-context understanding and better alignment with user preferences.",
+        model_type=ModelType.LLM,
+        model_series=QWEN3_SERIES
+    )
+)
+
+Model.register(
+    dict(
+        model_id = "Qwen3-30B-A3B-Thinking-2507",
+        supported_engines=[vllm_qwen3_engin091],
+        supported_instances=[
+            g5d12xlarge_instance,
+            g5d24xlarge_instance,
+            g5d48xlarge_instance,
+            local_instance
+        ],
+        supported_services=[
+            sagemaker_service,
+            sagemaker_async_service,
+            ecs_service,
+            local_service
+        ],
+        supported_frameworks=[
+            fastapi_framework
+        ],
+        allow_china_region=True,
+        huggingface_model_id="Qwen/Qwen3-30B-A3B-Thinking-2507",
+        modelscope_model_id="Qwen/Qwen3-30B-A3B-Thinking-2507",
+        require_huggingface_token=False,
+        application_scenario="Advanced reasoning, mathematics, science, coding, logical reasoning, complex problem solving",
+        description="Qwen3-30B-A3B-Thinking-2507 is an enhanced thinking-enabled version featuring significantly improved performance on reasoning tasks, including logical reasoning, mathematics, science, coding, and academic benchmarks. This model supports only thinking mode with enhanced 256K long-context understanding and markedly better general capabilities for highly complex reasoning tasks.",
+        model_type=ModelType.LLM,
+        model_series=QWEN3_SERIES
+    )
+)
+
 
 Model.register(
     dict(