fix(anthropic): support automatic compaction (Opus 4.6) (#35034)

libs/partners/anthropic/langchain_anthropic/chat_models.py

@@ -848,6 +848,9 @@ class ChatAnthropic(BaseChatModel):
     """Parameters for Claude reasoning,
 
     e.g., `#!python {"type": "enabled", "budget_tokens": 10_000}`
+
+    For Claude Opus 4.6, `budget_tokens` is deprecated in favor of
+    `#!python {"type": "adaptive"}`
     """
 
     effort: Literal["high", "medium", "low"] | None = None
@@ -896,6 +899,12 @@ class ChatAnthropic(BaseChatModel):
     invocations.
     """
 
+    inference_geo: str | None = None
+    """Controls where model inference runs. See Anthropic's
+    [data residency](https://platform.claude.com/docs/en/build-with-claude/data-residency)
+    docs for more information.
+    """
+
     @property
     def _llm_type(self) -> str:
         """Return type of chat model."""
@@ -1122,6 +1131,8 @@ class ChatAnthropic(BaseChatModel):
         }
         if self.thinking is not None:
             payload["thinking"] = self.thinking
+        if self.inference_geo is not None:
+            payload["inference_geo"] = self.inference_geo
 
         # Handle output_config and effort parameter
         # Priority: self.effort > payload output_config
@@ -1273,6 +1284,7 @@ class ChatAnthropic(BaseChatModel):
                 not _tools_in_params(payload)
                 and not _documents_in_params(payload)
                 and not _thinking_in_params(payload)
+                and not _compact_in_params(payload)
             )
             block_start_event = None
             for event in stream:
@@ -1309,6 +1321,7 @@ class ChatAnthropic(BaseChatModel):
                 not _tools_in_params(payload)
                 and not _documents_in_params(payload)
                 and not _thinking_in_params(payload)
+                and not _compact_in_params(payload)
             )
             block_start_event = None
             async for event in stream:
@@ -1870,6 +1883,12 @@ def _documents_in_params(params: dict) -> bool:
     return False
 
 
+def _compact_in_params(params: dict) -> bool:
+    edits = params.get("context_management", {}).get("edits") or []
+
+    return any("compact" in (edit.get("type") or "") for edit in edits)
+
+
 class _AnthropicToolUse(TypedDict):
     type: Literal["tool_use"]
     name: str
@@ -2049,6 +2068,13 @@ def _make_message_chunk_from_anthropic_event(
                 tool_call_chunks=tool_call_chunks,
             )
 
+        # Compaction block
+        elif event.delta.type == "compaction_delta":
+            content_block = event.delta.model_dump()
+            content_block["index"] = event.index
+            content_block["type"] = "compaction"
+            message_chunk = AIMessageChunk(content=[content_block])
+
     # Process final usage metadata and completion info
     elif event.type == "message_delta" and stream_usage:
         usage_metadata = _create_usage_metadata(event.usage)

libs/partners/anthropic/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
 version = "1.3.1"
 requires-python = ">=3.10.0,<4.0.0"
 dependencies = [
-    "anthropic>=0.75.0,<1.0.0",
+    "anthropic>=0.78.0,<1.0.0",
     "langchain-core>=1.2.6,<2.0.0",
     "pydantic>=2.7.4,<3.0.0",
 ]

libs/partners/anthropic/tests/integration_tests/test_chat_models.py

@@ -2409,3 +2409,112 @@ def test_fine_grained_tool_streaming() -> None:
     assert write_doc_block is not None
     assert write_doc_block["name"] == "write_document"
     assert "args" in write_doc_block
+
+
+@pytest.mark.vcr
+def test_compaction() -> None:
+    """Test the compation beta feature."""
+    llm = ChatAnthropic(
+        model="claude-opus-4-6",  # type: ignore[call-arg]
+        betas=["compact-2026-01-12"],
+        max_tokens=4096,
+        context_management={
+            "edits": [
+                {
+                    "type": "compact_20260112",
+                    "trigger": {"type": "input_tokens", "value": 50000},
+                    "pause_after_compaction": True,
+                }
+            ]
+        },
+    )
+
+    input_message = {
+        "role": "user",
+        "content": f"Generate a one-sentence summary of this:\n\n{'a' * 100000}",
+    }
+    messages: list = [input_message]
+
+    first_response = llm.invoke(messages)
+    messages.append(first_response)
+
+    second_message = {
+        "role": "user",
+        "content": f"Generate a one-sentence summary of this:\n\n{'b' * 100000}",
+    }
+    messages.append(second_message)
+
+    second_response = llm.invoke(messages)
+    messages.append(second_response)
+
+    content_blocks = second_response.content_blocks
+    compaction_block = next(
+        (block for block in content_blocks if block["type"] == "non_standard"),
+        None,
+    )
+    assert compaction_block
+    assert compaction_block["value"].get("type") == "compaction"
+
+    third_message = {
+        "role": "user",
+        "content": "What are we talking about?",
+    }
+    messages.append(third_message)
+    third_response = llm.invoke(messages)
+    content_blocks = third_response.content_blocks
+    assert [block["type"] for block in content_blocks] == ["text"]
+
+
+@pytest.mark.vcr
+def test_compaction_streaming() -> None:
+    """Test the compation beta feature."""
+    llm = ChatAnthropic(
+        model="claude-opus-4-6",  # type: ignore[call-arg]
+        betas=["compact-2026-01-12"],
+        max_tokens=4096,
+        context_management={
+            "edits": [
+                {
+                    "type": "compact_20260112",
+                    "trigger": {"type": "input_tokens", "value": 50000},
+                    "pause_after_compaction": False,
+                }
+            ]
+        },
+        streaming=True,
+    )
+
+    input_message = {
+        "role": "user",
+        "content": f"Generate a one-sentence summary of this:\n\n{'a' * 100000}",
+    }
+    messages: list = [input_message]
+
+    first_response = llm.invoke(messages)
+    messages.append(first_response)
+
+    second_message = {
+        "role": "user",
+        "content": f"Generate a one-sentence summary of this:\n\n{'b' * 100000}",
+    }
+    messages.append(second_message)
+
+    second_response = llm.invoke(messages)
+    messages.append(second_response)
+
+    content_blocks = second_response.content_blocks
+    compaction_block = next(
+        (block for block in content_blocks if block["type"] == "non_standard"),
+        None,
+    )
+    assert compaction_block
+    assert compaction_block["value"].get("type") == "compaction"
+
+    third_message = {
+        "role": "user",
+        "content": "What are we talking about?",
+    }
+    messages.append(third_message)
+    third_response = llm.invoke(messages)
+    content_blocks = third_response.content_blocks
+    assert [block["type"] for block in content_blocks] == ["text"]

libs/partners/anthropic/tests/unit_tests/test_chat_models.py

@@ -1564,6 +1564,13 @@ def test_context_management_in_payload() -> None:
     }
 
 
+def test_inference_geo_in_payload() -> None:
+    llm = ChatAnthropic(model=MODEL_NAME, inference_geo="us")
+    input_message = HumanMessage("Hello, world!")
+    payload = llm._get_request_payload([input_message])
+    assert payload["inference_geo"] == "us"
+
+
 def test_anthropic_model_params() -> None:
     llm = ChatAnthropic(model=MODEL_NAME)
 

libs/partners/anthropic/uv.lock

@@ -1,5 +1,5 @@
 version = 1
-revision = 3
+revision = 2
 requires-python = ">=3.10.0, <4.0.0"
 resolution-markers = [
     "python_full_version >= '3.13' and platform_python_implementation == 'PyPy'",
@@ -24,7 +24,7 @@ wheels = [
 
 [[package]]
 name = "anthropic"
-version = "0.75.0"
+version = "0.78.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "anyio" },
@@ -36,9 +36,9 @@ dependencies = [
     { name = "sniffio" },
     { name = "typing-extensions" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/04/1f/08e95f4b7e2d35205ae5dcbb4ae97e7d477fc521c275c02609e2931ece2d/anthropic-0.75.0.tar.gz", hash = "sha256:e8607422f4ab616db2ea5baacc215dd5f028da99ce2f022e33c7c535b29f3dfb", size = 439565, upload-time = "2025-11-24T20:41:45.28Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/ec/51/32849a48f9b1cfe80a508fd269b20bd8f0b1357c70ba092890fde5a6a10b/anthropic-0.78.0.tar.gz", hash = "sha256:55fd978ab9b049c61857463f4c4e9e092b24f892519c6d8078cee1713d8af06e", size = 509136, upload-time = "2026-02-05T17:52:04.986Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/60/1c/1cd02b7ae64302a6e06724bf80a96401d5313708651d277b1458504a1730/anthropic-0.75.0-py3-none-any.whl", hash = "sha256:ea8317271b6c15d80225a9f3c670152746e88805a7a61e14d4a374577164965b", size = 388164, upload-time = "2025-11-24T20:41:43.587Z" },
+    { url = "https://files.pythonhosted.org/packages/3b/03/2f50931a942e5e13f80e24d83406714672c57964be593fc046d81369335b/anthropic-0.78.0-py3-none-any.whl", hash = "sha256:2a9887d2e99d1b0f9fe08857a1e9fe5d2d4030455dbf9ac65aab052e2efaeac4", size = 405485, upload-time = "2026-02-05T17:52:03.674Z" },
 ]
 
 [[package]]
@@ -498,7 +498,7 @@ wheels = [
 
 [[package]]
 name = "langchain"
-version = "1.2.7"
+version = "1.2.9"
 source = { editable = "../../langchain_v1" }
 dependencies = [
     { name = "langchain-core" },
@@ -607,7 +607,7 @@ typing = [
 
 [package.metadata]
 requires-dist = [
-    { name = "anthropic", specifier = ">=0.75.0,<1.0.0" },
+    { name = "anthropic", specifier = ">=0.78.0,<1.0.0" },
     { name = "langchain-core", editable = "../../core" },
     { name = "pydantic", specifier = ">=2.7.4,<3.0.0" },
 ]
@@ -646,7 +646,7 @@ typing = [
 
 [[package]]
 name = "langchain-core"
-version = "1.2.7"
+version = "1.2.9"
 source = { editable = "../../core" }
 dependencies = [
     { name = "jsonpatch" },
@@ -706,7 +706,7 @@ typing = [
 
 [[package]]
 name = "langchain-tests"
-version = "1.1.2"
+version = "1.1.4"
 source = { editable = "../../standard-tests" }
 dependencies = [
     { name = "httpx" },