diff --git a/pyproject.toml b/pyproject.toml
index a4151e5..cf983a5 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -32,3 +32,10 @@ allow-direct-references = true
 [tool.black]
 line-length = 120
 skip-string-normalization = true
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+python_files = ["test_*.py"]
+python_classes = ["Test*"]
+python_functions = ["test_*"]
+addopts = "-v --tb=short"
diff --git a/spectrumlab/config/base_config.py b/spectrumlab/config/base_config.py
index 11d54aa..c0640c9 100644
--- a/spectrumlab/config/base_config.py
+++ b/spectrumlab/config/base_config.py
@@ -11,6 +11,10 @@
 
 @dataclass
 class Config:
+    # This api key is for testing closed MLLMs by Boyue Richdata
+    BOYUE_API_KEY: str = os.getenv("BOYUE_API_KEY")
+    BOYUE_BASE_URL: str = os.getenv("BOYUE_BASE_URL")
+
     # DeepSeek API Configuration
     deepseek_api_key: str = os.getenv("DEEPSEEK_API_KEY")
     deepseek_base_url: str = os.getenv("DEEPSEEK_BASE_URL")
@@ -25,3 +29,29 @@ class Config:
     internvl_api_key: str = os.getenv("INTERNVL_API_KEY")
     internvl_base_url: str = os.getenv("INTERNVL_BASE_URL")
     internvl_model_name: str = os.getenv("INTERNVL_MODEL_NAME")
+
+    # Claude API Configuration
+    claude_api_key: str = BOYUE_API_KEY
+    claude_base_url: str = BOYUE_BASE_URL
+    claude_sonnet_3_5_model_name: str = os.getenv("CLAUDE_SONNET_3_5")
+    claude_opus_4_model_name: str = os.getenv("CLAUDE_OPUS_4")
+    claude_haiku_3_5_model_name: str = os.getenv("CLAUDE_HAIKU_3_5")
+    claude_sonnet_4_model_name: str = os.getenv("CLAUDE_SONNET_4")
+
+    # GPT-4.1, GPT-4-Vision
+    gpt4_1_api_key: str = BOYUE_API_KEY
+    gpt4_1_base_url: str = BOYUE_BASE_URL
+    gpt4_1_model_name: str = os.getenv("GPT4_1")
+    gpt4_vision_api_key: str = BOYUE_API_KEY
+    gpt4_vision_base_url: str = BOYUE_BASE_URL
+    gpt4_vision_model_name: str = os.getenv("GPT4_VISION")
+
+    # Grok-2-Vision
+    grok_2_vision_api_key: str = BOYUE_API_KEY
+    grok_2_vision_base_url: str = BOYUE_BASE_URL
+    grok_2_vision_model_name: str = os.getenv("GROK_2_VISION")
+
+    # Qwen-VL-Max
+    qwen_vl_api_key: str = BOYUE_API_KEY
+    qwen_vl_base_url: str = BOYUE_BASE_URL
+    qwen_vl_model_name: str = os.getenv("QWEN_VL")
diff --git a/spectrumlab/models/__init__.py b/spectrumlab/models/__init__.py
index 93ce4d1..0750ae3 100644
--- a/spectrumlab/models/__init__.py
+++ b/spectrumlab/models/__init__.py
@@ -1,5 +1,10 @@
 from .deepseek_api import DeepSeek
 from .gpt4o_api import GPT4o
 from .internvl_api import InternVL
+from .claude_api import Claude_Sonnet_3_5, Claude_Opus_4, Claude_Haiku_3_5, Claude_Sonnet_4
+from .gpt4_v_api import GPT4_1, GPT4_Vision
+from .grok_api import Grok_2_Vision
+from .qwen_vl_api import Qwen_VL_Max
 
-__all__ = ["DeepSeek", "GPT4o", "InternVL"]
+__all__ = ["DeepSeek", "GPT4o", "InternVL", "Claude_Sonnet_3_5", "Claude_Opus_4", 
+           "Claude_Haiku_3_5", "Claude_Sonnet_4", "GPT4_1", "GPT4_Vision", "Grok_2_Vision", "Qwen_VL_Max"]
diff --git a/spectrumlab/models/claude_api.py b/spectrumlab/models/claude_api.py
new file mode 100644
index 0000000..9711061
--- /dev/null
+++ b/spectrumlab/models/claude_api.py
@@ -0,0 +1,288 @@
+from typing import Optional, Union, Dict, Any
+from .base_api import BaseAPIModel
+from spectrumlab.config import Config
+from openai import OpenAI
+
+
+class Claude_Sonnet_3_5(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.BOYUE_API_KEY
+        self.base_url = base_url or config.BOYUE_BASE_URL
+        self.model_name = model_name or config.claude_sonnet_3_5_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "Claude API key not found. Please set CLAUDE_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_out_len: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_out_len: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+
+        # Link: https://docs.anthropic.com/claude/reference/getting-started-with-the-api
+        messages = []
+
+        if isinstance(prompt, dict) and "images" in prompt:
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_out_len,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"Claude API call failed: {e}")
+
+
+class Claude_Opus_4(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.BOYUE_API_KEY
+        self.base_url = base_url or config.BOYUE_BASE_URL
+        self.model_name = model_name or config.claude_opus_4_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "Claude API key not found. Please set CLAUDE_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_out_len: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_out_len: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+
+        # Link: https://docs.anthropic.com/claude/reference/getting-started-with-the-api
+        messages = []
+
+        if isinstance(prompt, dict) and "images" in prompt:
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_out_len,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"Claude API call failed: {e}")
+
+
+class Claude_Haiku_3_5(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.BOYUE_API_KEY
+        self.base_url = base_url or config.BOYUE_BASE_URL
+        self.model_name = model_name or config.claude_haiku_3_5_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "Claude API key not found. Please set CLAUDE_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_out_len: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_out_len: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+
+        # Link: https://docs.anthropic.com/claude/reference/getting-started-with-the-api
+        messages = []
+
+        if isinstance(prompt, dict) and "images" in prompt:
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_out_len,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"Claude API call failed: {e}")
+
+class Claude_Sonnet_4(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.BOYUE_API_KEY
+        self.base_url = base_url or config.BOYUE_BASE_URL
+        self.model_name = model_name or config.claude_sonnet_4_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "Claude API key not found. Please set CLAUDE_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_out_len: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_out_len: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+
+        # Link: https://docs.anthropic.com/claude/reference/getting-started-with-the-api
+        messages = []
+
+        if isinstance(prompt, dict) and "images" in prompt:
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_out_len,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"Claude API call failed: {e}")
+
diff --git a/spectrumlab/models/gpt4_v_api.py b/spectrumlab/models/gpt4_v_api.py
new file mode 100644
index 0000000..c14312c
--- /dev/null
+++ b/spectrumlab/models/gpt4_v_api.py
@@ -0,0 +1,147 @@
+from typing import Dict, Any, Optional, Union
+from .base_api import BaseAPIModel
+from spectrumlab.config import Config
+from openai import OpenAI
+
+
+class GPT4_1(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.gpt4_1_api_key
+        self.base_url = base_url or config.gpt4_1_base_url
+        self.model_name = model_name or config.gpt4_1_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "GPT-4.1 API key not found. Please set GPT4_1_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_tokens: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_tokens: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+        messages = []
+
+        # Handle multimodal vs text-only prompts
+        if isinstance(prompt, dict) and "images" in prompt:
+            # Multimodal prompt
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            # Text-only prompt
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_tokens,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"GPT-4.1 API call failed: {e}")
+
+class GPT4_Vision(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.gpt4_vision_api_key
+        self.base_url = base_url or config.gpt4_vision_base_url
+        self.model_name = model_name or config.gpt4_vision_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "GPT-4 Vision API key not found. Please set GPT4_VISION_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_tokens: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_tokens: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+        messages = []
+
+        # Handle multimodal vs text-only prompts
+        if isinstance(prompt, dict) and "images" in prompt:
+            # Multimodal prompt
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            # Text-only prompt
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_tokens,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"GPT-4 Vision API call failed: {e}")
\ No newline at end of file
diff --git a/spectrumlab/models/grok_api.py b/spectrumlab/models/grok_api.py
new file mode 100644
index 0000000..243287a
--- /dev/null
+++ b/spectrumlab/models/grok_api.py
@@ -0,0 +1,76 @@
+from typing import Dict, Any, Optional, Union
+from .base_api import BaseAPIModel
+from spectrumlab.config import Config
+from openai import OpenAI
+
+
+class Grok_2_Vision(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.grok_2_vision_api_key
+        self.base_url = base_url or config.grok_2_vision_base_url
+        self.model_name = model_name or config.grok_2_vision_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "Grok-2-Vision API key not found. Please set GROK_2_VISION_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_tokens: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_tokens: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+        messages = []
+
+        # Handle multimodal vs text-only prompts
+        if isinstance(prompt, dict) and "images" in prompt:
+            # Multimodal prompt
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            # Text-only prompt
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_tokens,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"Grok-2-Vision API call failed: {e}")
diff --git a/spectrumlab/models/qwen_vl_api.py b/spectrumlab/models/qwen_vl_api.py
new file mode 100644
index 0000000..d4c6817
--- /dev/null
+++ b/spectrumlab/models/qwen_vl_api.py
@@ -0,0 +1,76 @@
+from typing import Dict, Any, Optional, Union
+from .base_api import BaseAPIModel
+from spectrumlab.config import Config
+from openai import OpenAI
+
+
+class Qwen_VL_Max(BaseAPIModel):
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        model_name: Optional[str] = None,
+        **kwargs,
+    ):
+        config = Config()
+
+        # Use provided parameters or fall back to config
+        self.api_key = api_key or config.qwen_vl_api_key
+        self.base_url = base_url or config.qwen_vl_base_url
+        self.model_name = model_name or config.qwen_vl_model_name
+
+        # Validate that we have required configuration
+        if not self.api_key:
+            raise ValueError(
+                "Qwen-VL-Max API key not found. Please set QWEN_VL_MAX_API_KEY in your .env file "
+                "or provide api_key parameter."
+            )
+
+        self.client = OpenAI(
+            api_key=self.api_key,
+            base_url=self.base_url,
+        )
+
+        # Initialize parent class
+        super().__init__(model_name=self.model_name, **kwargs)
+
+    def generate(
+        self, prompt: Union[str, Dict[str, Any]], max_tokens: int = 512
+    ) -> str:
+        """
+        Generate response supporting both text and multimodal input.
+
+        Args:
+            prompt: Either text string or multimodal dict
+            max_tokens: Maximum tokens to generate
+
+        Returns:
+            Generated response string
+        """
+        messages = []
+
+        # Handle multimodal vs text-only prompts
+        if isinstance(prompt, dict) and "images" in prompt:
+            # Multimodal prompt
+            content = []
+
+            content.append({"type": "text", "text": prompt["text"]})
+
+            for image_data in prompt["images"]:
+                content.append(image_data)
+
+            messages.append({"role": "user", "content": content})
+        else:
+            # Text-only prompt
+            text_content = prompt if isinstance(prompt, str) else prompt.get("text", "")
+            messages.append({"role": "user", "content": text_content})
+
+        try:
+            response = self.client.chat.completions.create(
+                model=self.model_name,
+                messages=messages,
+                max_tokens=max_tokens,
+            )
+            return response.choices[0].message.content
+        except Exception as e:
+            raise RuntimeError(f"Qwen-VL-Max API call failed: {e}")
diff --git a/tests/models/test_claude_haiku_3_5.py b/tests/models/test_claude_haiku_3_5.py
new file mode 100644
index 0000000..16ea288
--- /dev/null
+++ b/tests/models/test_claude_haiku_3_5.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import Claude_Haiku_3_5
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = Claude_Haiku_3_5()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = Claude_Haiku_3_5()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = Claude_Haiku_3_5()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_claude_opus_4.py b/tests/models/test_claude_opus_4.py
new file mode 100644
index 0000000..66b383b
--- /dev/null
+++ b/tests/models/test_claude_opus_4.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import Claude_Opus_4
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = Claude_Opus_4()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = Claude_Opus_4()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = Claude_Opus_4()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_claude_sonnet_3_5.py b/tests/models/test_claude_sonnet_3_5.py
new file mode 100644
index 0000000..818470a
--- /dev/null
+++ b/tests/models/test_claude_sonnet_3_5.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import Claude_Sonnet_3_5
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = Claude_Sonnet_3_5()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = Claude_Sonnet_3_5()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = Claude_Sonnet_3_5()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_claude_sonnet_4.py b/tests/models/test_claude_sonnet_4.py
new file mode 100644
index 0000000..1d284b3
--- /dev/null
+++ b/tests/models/test_claude_sonnet_4.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import Claude_Sonnet_4
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = Claude_Sonnet_4()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = Claude_Sonnet_4()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = Claude_Sonnet_4()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_gpt_4_1.py b/tests/models/test_gpt_4_1.py
new file mode 100644
index 0000000..1982848
--- /dev/null
+++ b/tests/models/test_gpt_4_1.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import GPT4_1
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = GPT4_1()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = GPT4_1()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = GPT4_1()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_gpt_4_v.py b/tests/models/test_gpt_4_v.py
new file mode 100644
index 0000000..9b52442
--- /dev/null
+++ b/tests/models/test_gpt_4_v.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import GPT4_Vision
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = GPT4_Vision()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = GPT4_Vision()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = GPT4_Vision()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_grok_2_v.py b/tests/models/test_grok_2_v.py
new file mode 100644
index 0000000..29bb571
--- /dev/null
+++ b/tests/models/test_grok_2_v.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import Grok_2_Vision
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = Grok_2_Vision()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = Grok_2_Vision()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = Grok_2_Vision()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]
diff --git a/tests/models/test_internvl.py b/tests/models/test_internvl.py
index f718961..f469c90 100644
--- a/tests/models/test_internvl.py
+++ b/tests/models/test_internvl.py
@@ -14,14 +14,14 @@ def test_internvl_text_generation():
 
 def test_internvl_multimodal_generation():
     model = InternVL()
-    image_path = "playground/models/test.png"
+    image_path = "playground/models/test.jpg"
     image_base64 = encode_image_to_base64(image_path)
     prompt = {
         "text": "Please explain this spectroscopy image.",
         "images": [
             {
                 "type": "image_url",
-                "image_url": {"url": f"data:image/png;base64,{image_base64}"},
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
             }
         ],
     }
diff --git a/tests/models/test_qwen_vl.py b/tests/models/test_qwen_vl.py
new file mode 100644
index 0000000..8310935
--- /dev/null
+++ b/tests/models/test_qwen_vl.py
@@ -0,0 +1,40 @@
+from spectrumlab.models import Qwen_VL_Max
+from spectrumlab.utils.image_utils import encode_image_to_base64
+from spectrumlab.benchmark.signal_group import SignalGroup
+from spectrumlab.evaluator.choice_evaluator import ChoiceEvaluator
+
+
+def test_claude_text_generation():
+    model = Qwen_VL_Max()
+    prompt = "What is spectroscopy?"
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_multimodal_generation():
+    model = Qwen_VL_Max()
+    image_path = "playground/models/test.jpg"
+    image_base64 = encode_image_to_base64(image_path)
+    prompt = {
+        "text": "Please explain this spectroscopy image.",
+        "images": [
+            {
+                "type": "image_url",
+                "image_url": {"url": f"data:image/jpg;base64,{image_base64}"},
+            }
+        ],
+    }
+    response = model.generate(prompt)
+    assert isinstance(response, str)
+    assert len(response) > 0
+
+
+def test_claude_signalgroup_evaluation():
+    model = Qwen_VL_Max()
+    signal_group = SignalGroup("data")
+    data = signal_group.get_data_by_subcategories(["Spectrum Type Classification"])
+    evaluator = ChoiceEvaluator()
+    results = evaluator.evaluate(data_items=data, model=model, save_path=None)
+    assert "metrics" in results
+    assert "overall" in results["metrics"]