Fix: Add dimensions parameter support for text-embedding-v4 model

libs/agno/agno/knowledge/embedder/azure_openai.py

@@ -100,7 +100,7 @@ def _response(self, text: str) -> CreateEmbeddingResponse:
         }
         if self.user is not None:
             _request_params["user"] = self.user
-        if self.id.startswith("text-embedding-3"):
+        if self.dimensions is not None:
             _request_params["dimensions"] = self.dimensions
         if self.request_params:
             _request_params.update(self.request_params)
@@ -131,7 +131,7 @@ async def _aresponse(self, text: str) -> CreateEmbeddingResponse:
         }
         if self.user is not None:
             _request_params["user"] = self.user
-        if self.id.startswith("text-embedding-3"):
+        if self.dimensions is not None:
             _request_params["dimensions"] = self.dimensions
         if self.request_params:
             _request_params.update(self.request_params)
@@ -181,7 +181,7 @@ async def async_get_embeddings_batch_and_usage(
             }
             if self.user is not None:
                 req["user"] = self.user
-            if self.id.startswith("text-embedding-3"):
+            if self.dimensions is not None:
                 req["dimensions"] = self.dimensions
             if self.request_params:
                 req.update(self.request_params)

libs/agno/agno/knowledge/embedder/openai.py

@@ -71,7 +71,7 @@ def response(self, text: str) -> CreateEmbeddingResponse:
         }
         if self.user is not None:
             _request_params["user"] = self.user
-        if self.id.startswith("text-embedding-3"):
+        if self.dimensions is not None:
             _request_params["dimensions"] = self.dimensions
         if self.request_params:
             _request_params.update(self.request_params)
@@ -106,7 +106,7 @@ async def async_get_embedding(self, text: str) -> List[float]:
         }
         if self.user is not None:
             req["user"] = self.user
-        if self.id.startswith("text-embedding-3"):
+        if self.dimensions is not None:
             req["dimensions"] = self.dimensions
         if self.request_params:
             req.update(self.request_params)
@@ -126,7 +126,7 @@ async def async_get_embedding_and_usage(self, text: str):
         }
         if self.user is not None:
             req["user"] = self.user
-        if self.id.startswith("text-embedding-3"):
+        if self.dimensions is not None:
             req["dimensions"] = self.dimensions
         if self.request_params:
             req.update(self.request_params)
@@ -166,7 +166,7 @@ async def async_get_embeddings_batch_and_usage(
             }
             if self.user is not None:
                 req["user"] = self.user
-            if self.id.startswith("text-embedding-3"):
+            if self.dimensions is not None:
                 req["dimensions"] = self.dimensions
             if self.request_params:
                 req.update(self.request_params)

libs/agno/tests/unit/embedder/test_openai_dimensions.py

@@ -0,0 +1,106 @@
+import importlib.util
+from types import SimpleNamespace
+
+import pytest
+
+from agno.knowledge.embedder.openai import OpenAIEmbedder
+
+
+@pytest.mark.skipif(not importlib.util.find_spec("openai"), reason="openai package not installed")
+def test_dimensions_propagated_when_explicitly_set():
+    """Ensure dimensions parameter is passed when explicitly set, regardless of model."""
+
+    class DummyEmbeddings:
+        def __init__(self):
+            self.last_kwargs = None
+
+        def create(self, **kwargs):
+            self.last_kwargs = kwargs
+            dims = kwargs.get("dimensions", 1)
+            return SimpleNamespace(
+                data=[SimpleNamespace(embedding=[0.0] * dims)],
+                usage=None,
+            )
+
+    class DummyClient:
+        def __init__(self):
+            self.embeddings = DummyEmbeddings()
+
+    embedder = OpenAIEmbedder(id="text-embedding-v4", dimensions=512)
+    embedder.openai_client = DummyClient()
+
+    _ = embedder.get_embedding("hello world")
+
+    assert embedder.openai_client.embeddings.last_kwargs is not None, "Embeddings request not captured"
+    assert (
+        embedder.openai_client.embeddings.last_kwargs.get("dimensions") == 512
+    ), "dimensions parameter not propagated when explicitly set"
+
+
+@pytest.mark.skipif(not importlib.util.find_spec("openai"), reason="openai package not installed")
+def test_dimensions_propagated_for_any_model():
+    """Ensure dimensions parameter is passed for ANY model when explicitly set (future-proof)."""
+
+    class DummyEmbeddings:
+        def __init__(self):
+            self.last_kwargs = None
+
+        def create(self, **kwargs):
+            self.last_kwargs = kwargs
+            dims = kwargs.get("dimensions", 1536)
+            return SimpleNamespace(
+                data=[SimpleNamespace(embedding=[0.0] * dims)],
+                usage=None,
+            )
+
+    class DummyClient:
+        def __init__(self):
+            self.embeddings = DummyEmbeddings()
+
+    embedder = OpenAIEmbedder(id="text-embedding-ada-002", dimensions=256)
+    embedder.openai_client = DummyClient()
+
+    _ = embedder.get_embedding("test")
+
+    assert embedder.openai_client.embeddings.last_kwargs.get("dimensions") == 256, (
+        "dimensions should be passed for legacy models too"
+    )
+
+    embedder2 = OpenAIEmbedder(id="text-embedding-v5-ultra", dimensions=2048)
+    embedder2.openai_client = DummyClient()
+
+    _ = embedder2.get_embedding("test")
+
+    assert embedder2.openai_client.embeddings.last_kwargs.get("dimensions") == 2048, (
+        "dimensions should be passed for any future models"
+    )
+
+
+@pytest.mark.skipif(not importlib.util.find_spec("openai"), reason="openai package not installed")
+def test_dimensions_not_passed_when_none():
+    """Ensure dimensions parameter is NOT passed when set to None (respects model defaults)."""
+
+    class DummyEmbeddings:
+        def __init__(self):
+            self.last_kwargs = None
+
+        def create(self, **kwargs):
+            self.last_kwargs = kwargs
+            return SimpleNamespace(
+                data=[SimpleNamespace(embedding=[0.0] * 1536)],
+                usage=None,
+            )
+
+    class DummyClient:
+        def __init__(self):
+            self.embeddings = DummyEmbeddings()
+
+    embedder = OpenAIEmbedder(id="text-embedding-3-small")
+    embedder.dimensions = None
+    embedder.openai_client = DummyClient()
+
+    _ = embedder.get_embedding("test")
+
+    assert "dimensions" not in embedder.openai_client.embeddings.last_kwargs, (
+        "dimensions should NOT be passed when None (use model default)"
+    )