Add gpt-4o-mini model support

Author: ajhai

llmstack/apps/types/agent.py

@@ -15,6 +15,7 @@ class AgentModel(str, Enum):
     GPT_4_LATEST = "gpt-4-turbo-latest"
     GPT_4_1106_PREVIEW = "gpt-4-1106-preview"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
 
     def __str__(self):
         return self.value
@@ -23,7 +24,7 @@ def __str__(self):
 class AgentConfigSchema(BaseSchema):
     model: AgentModel = Field(
         title="Model",
-        default=AgentModel.GPT_3_5_LATEST,
+        default=AgentModel.GPT_4_O_MINI,
         description="The model to use for the agent.",
     )
     user_message: Optional[str] = Field(

llmstack/common/blocks/llm/openai.py

@@ -292,6 +292,7 @@ class OpenAIChatCompletionsAPIProcessorOutput(OpenAIAPIProcessorOutput):
 class ChatCompletionsModel(str, Enum):
     GPT_4 = "gpt-4"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
     GPT_4_32K = "gpt-4-32k"
     GPT_4_TURBO = "gpt-4-turbo"
     GPT_3_5 = "gpt-3.5-turbo"
@@ -307,7 +308,7 @@ class OpenAIChatCompletionsAPIProcessorConfiguration(
     OpenAIAPIProcessorConfiguration,
 ):
     model: ChatCompletionsModel = Field(
-        default=ChatCompletionsModel.GPT_3_5,
+        default=ChatCompletionsModel.GPT_4_O_MINI,
         description="ID of the model to use. Currently, only `gpt-3.5-turbo` and `gpt-4` are supported.",
     )
     max_tokens: Optional[conint(ge=1, le=32000)] = Field(

llmstack/common/utils/sslr/resources/chat/completions.py

@@ -66,6 +66,7 @@ def create(
                 "gpt-4-vision-preview",
                 "gpt-4",
                 "gpt-4o",
+                "gpt-4o-mini",
                 "gpt-4-0314",
                 "gpt-4-0613",
                 "gpt-4-32k",

llmstack/processors/providers/azure/azure_chat_completions.py

@@ -16,6 +16,7 @@
 class ChatCompletionsModel(str, Enum):
     GPT_4 = "gpt-4"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
     GPT_4_TURBO = "gpt-4-turbo"
     GPT_3_5 = "gpt-35-turbo"
     GPT_3_5_16 = "gpt-35-turbo-16k"

llmstack/processors/providers/openai/chat_completions.py

@@ -24,6 +24,7 @@
 class ChatCompletionsModel(str, Enum):
     GPT_4 = "gpt-4"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
     GPT_4_32K = "gpt-4-32k"
     GPT_4_TURBO = "gpt-4-turbo"
     GPT_3_5 = "gpt-3.5-turbo"
@@ -128,7 +129,7 @@ class ChatCompletionsConfiguration(
     ApiProcessorSchema,
 ):
     model: str = Field(
-        default=ChatCompletionsModel.GPT_3_5,
+        default=ChatCompletionsModel.GPT_4_O_MINI,
         description="ID of the model to use. Currently, only `gpt-3.5-turbo` and `gpt-4` are supported.",
         json_schema_extra={
             "advanced_parameter": False,

llmstack/processors/providers/openai/chat_completions_vision.py

@@ -33,6 +33,7 @@ class ChatCompletionsVisionModel(str, Enum):
     GPT_4_TURBO_240409 = "gpt-4-turbo-2024-04-09"
     GPT_4_1106_VISION_PREVIEW = "gpt-4-1106-vision-preview"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
 
     def __str__(self):
         return self.value

llmstack/processors/providers/promptly/chat_completions.py

@@ -26,6 +26,7 @@ class Model(str, Enum):
     GPT_3_5_TURBO = "gpt-3.5-turbo"
     GPT_4 = "gpt-4"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
     GPT_4_TURBO_PREVIEW = "gpt-4-turbo-preview"
     GEMINI_PRO = "gemini-pro"
     CLAUDE_2_1 = "claude-2.1"
@@ -66,6 +67,7 @@ class OpenAIModel(str, Enum):
     GPT_4 = "gpt-4"
     GPT_4_TURBO_PREVIEW = "gpt-4-turbo-preview"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
 
     def __str__(self):
         return self.value
@@ -76,7 +78,7 @@ def model_name(self):
 
 class OpenAIModelConfig(BaseModel):
     provider: Literal["openai"] = "openai"
-    model: OpenAIModel = Field(default=OpenAIModel.GPT_3_5_TURBO, description="The model for the LLM")
+    model: OpenAIModel = Field(default=OpenAIModel.GPT_4_O_MINI, description="The model for the LLM")
 
 
 class GoogleModel(str, Enum):

llmstack/processors/providers/promptly/text_chat.py

@@ -22,6 +22,7 @@
 class TextChatCompletionsModel(str, Enum):
     GPT_4 = "gpt-4"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
     GPT_4_LATEST = "gpt-4-turbo-latest"
     GPT_3_5 = "gpt-3.5-turbo"
     GPT_3_5_LATEST = "gpt-3.5-turbo-latest"
@@ -35,7 +36,7 @@ def __str__(self):
 
 class TextChatConfiguration(ApiProcessorSchema):
     model: TextChatCompletionsModel = Field(
-        default=TextChatCompletionsModel.GPT_3_5,
+        default=TextChatCompletionsModel.GPT_4_O_MINI,
         description="ID of the model to use. Currently, only `gpt-3.5-turbo` and `gpt-4` are supported.",
         json_schema_extra={"widget": "customselect"},
     )
@@ -88,11 +89,6 @@ class TextChatConfiguration(ApiProcessorSchema):
         default=True,
         description="Use Azure if available. Will fallback to OpenAI when unchecked",
     )
-    use_localai_if_available: bool = Field(
-        title="Use LocalAI if available",
-        default=False,
-        description="Use LocalAI if available. Will fallback to OpenAI or Azure OpenAI when unchecked",
-    )
     chat_history_in_doc_search: int = Field(
         title="Chat history in doc search",
         default=0,
@@ -289,19 +285,26 @@ def process(self) -> dict:
         elif model == "gpt-4-turbo-latest":
             model = "gpt-4-0125-preview"
 
+        # Check if azure is available
+        provider_config = None
         if self._config.use_azure_if_available:
+            try:
+                provider_config = self.get_provider_config(
+                    provider_slug="azure",
+                    processor_slug="*",
+                    model_slug=model,
+                )
+            except Exception:
+                pass
+
+        if self._config.use_azure_if_available and provider_config:
             if model == "gpt-3.5-turbo":
                 model = "gpt-35-turbo"
             elif model == "gpt-3.5-turbo-16k":
                 model = "gpt-35-turbo-16k"
             elif model == "gpt-3.5-turbo-latest":
                 model = "gpt-35-turbo-1106"
 
-            provider_config = self.get_provider_config(
-                provider_slug="azure",
-                processor_slug="*",
-                model_slug=model,
-            )
             openai_client = AzureOpenAI(
                 api_key=provider_config.api_key,
                 api_version=provider_config.api_version,

llmstack/processors/providers/promptly/web_browser.py

@@ -76,6 +76,7 @@ class Model(str, Enum):
     GPT_3_5_16K = "gpt-3.5-turbo-16k"
     GPT_4 = "gpt-4"
     GPT_4_O = "gpt-4o"
+    GPT_4_O_MINI = "gpt-4o-mini"
     GPT_4_32K = "gpt-4-32k"
     GPT_4_LATEST = "gpt-4-turbo-latest"
     GPT_4_V_LATEST = "gpt-4-vision-latest"