🌿 Fern Regeneration -- August 29, 2025

pyproject.toml

@@ -3,7 +3,7 @@ name = "cohere"
 
 [tool.poetry]
 name = "cohere"
-version = "5.17.0"
+version = "6.0.0"
 description = ""
 readme = "README.md"
 authors = []

src/cohere/__init__.py

@@ -61,6 +61,7 @@
     ChatTextGenerationEvent,
     ChatTextResponseFormat,
     ChatTextResponseFormatV2,
+    ChatThinkingContent,
     ChatToolCallDeltaEvent,
     ChatToolCallDeltaEventDelta,
     ChatToolCallDeltaEventDeltaMessage,
@@ -364,6 +365,7 @@
     "ChatTextGenerationEvent",
     "ChatTextResponseFormat",
     "ChatTextResponseFormatV2",
+    "ChatThinkingContent",
     "ChatToolCallDeltaEvent",
     "ChatToolCallDeltaEventDelta",
     "ChatToolCallDeltaEventDeltaMessage",

src/cohere/core/client_wrapper.py

@@ -22,10 +22,10 @@ def __init__(
 
     def get_headers(self) -> typing.Dict[str, str]:
         headers: typing.Dict[str, str] = {
-            "User-Agent": "cohere/5.17.0",
+            "User-Agent": "cohere/6.0.0",
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "cohere",
-            "X-Fern-SDK-Version": "5.17.0",
+            "X-Fern-SDK-Version": "6.0.0",
         }
         if self._client_name is not None:
             headers["X-Client-Name"] = self._client_name

src/cohere/types/__init__.py

@@ -73,6 +73,7 @@
 from .chat_text_generation_event import ChatTextGenerationEvent
 from .chat_text_response_format import ChatTextResponseFormat
 from .chat_text_response_format_v2 import ChatTextResponseFormatV2
+from .chat_thinking_content import ChatThinkingContent
 from .chat_tool_call_delta_event import ChatToolCallDeltaEvent
 from .chat_tool_call_delta_event_delta import ChatToolCallDeltaEventDelta
 from .chat_tool_call_delta_event_delta_message import ChatToolCallDeltaEventDeltaMessage
@@ -289,6 +290,7 @@
     "ChatTextGenerationEvent",
     "ChatTextResponseFormat",
     "ChatTextResponseFormatV2",
+    "ChatThinkingContent",
     "ChatToolCallDeltaEvent",
     "ChatToolCallDeltaEventDelta",
     "ChatToolCallDeltaEventDeltaMessage",

src/cohere/types/assistant_message_response_content_item.py

@@ -24,13 +24,16 @@ class Config:
 
 
 class ThinkingAssistantMessageResponseContentItem(UncheckedBaseModel):
-    value: typing.Optional[typing.Any] = None
     type: typing.Literal["thinking"] = "thinking"
+    thinking: str
 
-    if not IS_PYDANTIC_V2:
+    if IS_PYDANTIC_V2:
+        model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow")  # type: ignore # Pydantic v2
+    else:
 
         class Config:
             smart_union = True
+            extra = pydantic.Extra.allow
 
 
 AssistantMessageResponseContentItem = typing_extensions.Annotated[

src/cohere/types/assistant_message_v2content_item.py

@@ -24,13 +24,16 @@ class Config:
 
 
 class ThinkingAssistantMessageV2ContentItem(UncheckedBaseModel):
-    value: typing.Optional[typing.Any] = None
     type: typing.Literal["thinking"] = "thinking"
+    thinking: str
 
-    if not IS_PYDANTIC_V2:
+    if IS_PYDANTIC_V2:
+        model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow")  # type: ignore # Pydantic v2
+    else:
 
         class Config:
             smart_union = True
+            extra = pydantic.Extra.allow
 
 
 AssistantMessageV2ContentItem = typing_extensions.Annotated[

src/cohere/types/chat_content_start_event_delta_message_content.py

@@ -9,6 +9,7 @@
 
 
 class ChatContentStartEventDeltaMessageContent(UncheckedBaseModel):
+    thinking: typing.Optional[str] = None
     text: typing.Optional[str] = None
     type: typing.Optional[ChatContentStartEventDeltaMessageContentType] = None
 

src/cohere/types/chat_thinking_content.py

@@ -0,0 +1,23 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+import pydantic
+from ..core.pydantic_utilities import IS_PYDANTIC_V2
+from ..core.unchecked_base_model import UncheckedBaseModel
+
+
+class ChatThinkingContent(UncheckedBaseModel):
+    """
+    Thinking content of the message. This will be present when `thinking` is enabled, and will contain the models internal reasoning.
+    """
+
+    thinking: str
+
+    if IS_PYDANTIC_V2:
+        model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow")  # type: ignore # Pydantic v2
+    else:
+
+        class Config:
+            smart_union = True
+            extra = pydantic.Extra.allow

src/cohere/types/embed_by_type_response_embeddings.py

@@ -41,6 +41,11 @@ class EmbedByTypeResponseEmbeddings(UncheckedBaseModel):
     An array of packed unsigned binary embeddings. The length of each binary embedding is 1/8 the length of the float embeddings of the provided model. Each value is between 0 and 255.
     """
 
+    base64: typing.Optional[typing.List[str]] = pydantic.Field(default=None)
+    """
+    An array of base64 embeddings. Each string is the result of appending the float embedding bytes together and base64 encoding that.
+    """
+
     if IS_PYDANTIC_V2:
         model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow")  # type: ignore # Pydantic v2
     else:

src/cohere/types/embedding_type.py

@@ -2,4 +2,4 @@
 
 import typing
 
-EmbeddingType = typing.Union[typing.Literal["float", "int8", "uint8", "binary", "ubinary"], typing.Any]
+EmbeddingType = typing.Union[typing.Literal["float", "int8", "uint8", "binary", "ubinary", "base64"], typing.Any]

src/cohere/types/thinking.py

@@ -10,12 +10,14 @@
 
 class Thinking(UncheckedBaseModel):
     """
-    Thinking gives the model enhanced reasoning capabilities for complex tasks, while also providing transparency into its step-by-step thought process before it delivers its final answer.
-    When thinking is turned on, the model creates thinking content blocks where it outputs its internal reasoning. The model will incorporate insights from this reasoning before crafting a final response.
-    When thinking is used without tools, the API response will include both thinking and text content blocks. Meanwhile, when thinking is used alongside tools and the model makes tool calls, the API response will include the thinking content block and `tool_calls`.
+    Configuration for [reasoning features](https://docs.cohere.com/docs/reasoning).
+    """
+
+    type: ThinkingType = pydantic.Field()
+    """
+    Reasoning is enabled by default for models that support it, but can be turned off by setting `"type": "disabled"`.
     """
 
-    type: ThinkingType
     token_budget: typing.Optional[int] = pydantic.Field(default=None)
     """
     The maximum number of tokens the model can use for thinking, which must be set to a positive integer.

src/cohere/v2/client.py

@@ -444,6 +444,7 @@ def embed(
             * `"uint8"`: Use this when you want to get back unsigned int8 embeddings. Supported with Embed v3.0 and newer Embed models.
             * `"binary"`: Use this when you want to get back signed binary embeddings. Supported with Embed v3.0 and newer Embed models.
             * `"ubinary"`: Use this when you want to get back unsigned binary embeddings. Supported with Embed v3.0 and newer Embed models.
+            * `"base64"`: Use this when you want to get back base64 embeddings. Supported with Embed v3.0 and newer Embed models.
 
         truncate : typing.Optional[V2EmbedRequestTruncate]
             One of `NONE|START|END` to specify how the API will handle inputs longer than the maximum token length.
@@ -996,6 +997,7 @@ async def embed(
             * `"uint8"`: Use this when you want to get back unsigned int8 embeddings. Supported with Embed v3.0 and newer Embed models.
             * `"binary"`: Use this when you want to get back signed binary embeddings. Supported with Embed v3.0 and newer Embed models.
             * `"ubinary"`: Use this when you want to get back unsigned binary embeddings. Supported with Embed v3.0 and newer Embed models.
+            * `"base64"`: Use this when you want to get back base64 embeddings. Supported with Embed v3.0 and newer Embed models.
 
         truncate : typing.Optional[V2EmbedRequestTruncate]
             One of `NONE|START|END` to specify how the API will handle inputs longer than the maximum token length.