feat(fireworks): Responses API for Fireworks AI

src/any_llm/providers/fireworks/fireworks.py

@@ -11,7 +11,9 @@
 from any_llm.provider import Provider
 from any_llm.providers.helpers import create_completion_from_response
 from any_llm.providers.fireworks.utils import _create_openai_chunk_from_fireworks_chunk
+from any_llm.types.responses import Response, ResponseStreamEvent
 
+from openai import Stream,OpenAI
 
 class FireworksProvider(Provider):
     PROVIDER_NAME = "Fireworks"
@@ -20,7 +22,7 @@ class FireworksProvider(Provider):
 
     SUPPORTS_COMPLETION_STREAMING = True
     SUPPORTS_COMPLETION = True
-    SUPPORTS_RESPONSES = False
+    SUPPORTS_RESPONSES = True
     SUPPORTS_COMPLETION_REASONING = False
     SUPPORTS_EMBEDDING = False
 
@@ -74,3 +76,23 @@ def _make_api_call(
             provider_name="Fireworks",
             model=model,
         )
+
+    def responses(self, model, input_data, **kwargs) -> Response | Iterator[ResponseStreamEvent]:
+        """Call Fireworks Responses API and normalize into ChatCompletion/Chunks."""
+        client = OpenAI(
+            base_url="https://api.fireworks.ai/inference/v1",
+            api_key=self.config.api_key,
+        )
+
+        response = client.responses.create(
+            model=model,
+            input=input_data,
+            **kwargs,
+        )
+
+        if not isinstance(response, (Response, Stream)):
+            raise ValueError(f"Responses API returned an unexpected type: {type(response)}")
+
+        response = Response.model_validate(response.model_dump())
+
+        return response

src/any_llm/providers/fireworks/utils.py

@@ -53,4 +53,4 @@ def _create_openai_chunk_from_fireworks_chunk(fireworks_chunk: Any) -> ChatCompl
         model=getattr(fireworks_chunk, "model", "unknown"),
         object="chat.completion.chunk",
         usage=usage,
-    )
+    )