Skip to content

strands.models.anthropic

Anthropic Claude model provider.

Beta that enables on-demand compaction, and compaction blocks in later requests.

Beta that lets a mid-conversation system message carry clear_at.

See https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages

class AnthropicModel(Model)

Defined in: src/strands/models/anthropic.py:223

Anthropic model provider implementation.

class AnthropicConfig(BaseModelConfig)

Defined in: src/strands/models/anthropic.py:243

Configuration options for Anthropic models.

Attributes:

  • cache_config - Configuration for prompt caching. Adds a cache point to the last user message, caching everything before it. Caching is off when unset.
  • cache_tools - Caches the tool definitions (deprecated, use CacheConfig(tools_ttl=…)). Superseded by an explicitly set cache_config.tools_ttl.
  • max_tokens - Maximum number of tokens to generate.
  • model_id - Calude model ID (e.g., “claude-3-7-sonnet-latest”). For a complete list of supported models, see https://docs.anthropic.com/en/docs/about-claude/models/all-models.
  • params - Additional model parameters (e.g., metadata, stop_sequences). For a complete list of supported parameters, see https://docs.anthropic.com/en/api/messages. Sampling parameters (temperature, top_k, top_p) are rejected from API version 2026-10-21 on; pin 2023-06-01 with client_args={“default_headers”: {“anthropic-version”: “2023-06-01”}} (or the same default_headers on a client passed as client=).
  • anthropic_tools - Anthropic-specific server-side tools that are not function tools (e.g., web_search, web_fetch, code_execution). Appended alongside the agent’s function tools. Use the standard tools interface for function calling tools. For a complete list of supported tools, see https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/overview
  • use_native_token_count - Whether to use the native Anthropic count_tokens API. When True, count_tokens() calls the Anthropic API for accurate counts. When False (default), skips the API call and uses the local estimator.
def __init__(*,
client: "_AsyncClient | None" = None,
client_args: dict[str, Any] | None = None,
**model_config: Unpack[AnthropicConfig])

Defined in: src/strands/models/anthropic.py:278

Initialize provider instance.

Arguments:

  • client - A pre-configured Anthropic-compatible async client, e.g. anthropic.AsyncAnthropicBedrock(), anthropic.AsyncAnthropicBedrockMantle() (anthropic>=0.91) or anthropic.AsyncAnthropicVertex(). The model does not close it.
  • client_args - Arguments for the underlying Anthropic client (e.g., api_key). For a complete list of supported arguments, see https://docs.anthropic.com/en/api/client-sdks.
  • **model_config - Configuration options for the Anthropic model.

Raises:

  • ValueError - If both client and client_args are provided.
@override
def update_config(**model_config: Unpack[AnthropicConfig]) -> None

Defined in: src/strands/models/anthropic.py:314

Update the Anthropic model configuration with the provided arguments.

Arguments:

  • **model_config - Configuration overrides.
@property
@override
def model_info() -> ModelInfo

Defined in: src/strands/models/anthropic.py:330

Everything the SDK knows about the configured model, with temporary_messages on for supported Claude.

temporary_messages follows the model id, not the client, so it is also on through the Bedrock and Vertex clients. If those reject the turn-scoped system messages, set model_info=ModelInfo(capabilities=ModelCapabilities(temporary_messages=False)). See :attr:strands.models.Model.model_info.

@override
def get_config() -> AnthropicConfig

Defined in: src/strands/models/anthropic.py:346

Get the Anthropic model configuration.

Returns:

The Anthropic model configuration.

def format_request(
messages: Messages,
tool_specs: list[ToolSpec] | None = None,
system_prompt: str | None = None,
tool_choice: ToolChoice | None = None,
dynamic_trailing_blocks: int = 0,
*,
system_prompt_content: list[SystemContentBlock] | None = None
) -> dict[str, Any]

Defined in: src/strands/models/anthropic.py:710

Format an Anthropic streaming request.

Arguments:

  • messages - List of message objects to be processed by the model.
  • tool_specs - List of tool specifications to make available to the model.
  • system_prompt - Plain string system prompt. Ignored when system_prompt_content is provided.
  • tool_choice - Selection strategy for tool invocation.
  • dynamic_trailing_blocks - How many trailing blocks of the last user message are rebuilt on every call, so the cache point stays ahead of them.
  • system_prompt_content - Structured system prompt content blocks, which can carry a cache point.

Returns:

An Anthropic streaming request.

Raises:

  • TypeError - If a message contains a content block type that cannot be converted to an Anthropic-compatible format.
  • ValueError - If params sets a sampling parameter both directly and in extra_body, or sets one while the request uses API version 2026-10-21 or later.
def format_chunk(event: dict[str, Any]) -> StreamEvent

Defined in: src/strands/models/anthropic.py:1036

Format the Anthropic response events into standardized message chunks.

Arguments:

  • event - A response event from the Anthropic model.

Returns:

The formatted chunk.

Raises:

  • RuntimeError - If chunk_type is not recognized. This error should never be encountered as we control chunk_type in the stream method.
@override
async def count_tokens(
messages: Messages,
tool_specs: list[ToolSpec] | None = None,
system_prompt: str | None = None,
system_prompt_content: list[SystemContentBlock] | None = None) -> int

Defined in: src/strands/models/anthropic.py:1158

Count tokens using Anthropic’s native count_tokens API.

Uses the same message format as the Messages API to get accurate token counts directly from the Anthropic service.

Arguments:

  • messages - List of message objects to count tokens for.
  • tool_specs - List of tool specifications to include in the count.
  • system_prompt - Plain string system prompt. Ignored if system_prompt_content is provided.
  • system_prompt_content - Structured system prompt content blocks.

Returns:

Total input token count.

@override
async def stream(messages: Messages,
tool_specs: list[ToolSpec] | None = None,
system_prompt: str | None = None,
*,
tool_choice: ToolChoice | None = None,
system_prompt_content: list[SystemContentBlock] | None = None,
structured_output_model: type[BaseModel] | None = None,
**kwargs: Any) -> AsyncGenerator[StreamEvent, None]

Defined in: src/strands/models/anthropic.py:1211

Stream conversation with the Anthropic model.

Arguments:

  • messages - List of message objects to be processed by the model.
  • tool_specs - List of tool specifications to make available to the model.
  • system_prompt - Plain string system prompt. Ignored when system_prompt_content is provided.
  • tool_choice - Selection strategy for tool invocation.
  • system_prompt_content - Structured system prompt content blocks, which can carry a cache point.
  • structured_output_model - Output model the call must produce; the request is then sent with the strongest structured output mechanism the model supports.
  • **kwargs - Additional keyword arguments for future extensibility.

Yields:

Formatted message chunks from the model.

A server-side tool turn that Anthropic pauses is resumed with follow-up requests inside this call, so an error raised by one of those requests can surface after earlier chunks were already yielded.

Raises:

  • ContextWindowOverflowException - If the input exceeds the model’s context window.
  • ModelThrottledException - If the request is throttled by Anthropic.
  • RuntimeError - If a paused server-side tool turn is still paused after the continuation limit.
  • ValueError - If params sets a sampling parameter both directly and in extra_body, or sets one while the request uses API version 2026-10-21 or later.
@override
async def summarize(messages: Messages,
*,
tool_specs: list[ToolSpec] | None = None,
system_prompt: str | None = None,
system_prompt_content: list[SystemContentBlock]
| None = None,
instructions: str | None = None,
cancel_signal: threading.Event | None = None,
agent_metadata: "AgentMetadata | None" = None,
**kwargs: Any) -> SummarizeResult

Defined in: src/strands/models/anthropic.py:1414

Summarize messages with Anthropic on-demand compaction.

Sends the request unchanged under beta compact-2026-09-04, so the thinking kept after a summary of the live conversation stays valid. The returned compaction block is stored as a summary block holding its text and signature; later requests from a model with server_side_summary rebuild the compaction block from it and start from it, and requests without it send the summary text. Uses the base :meth:Model.summarize when model_info.capabilities.server_side_summary is not True, when the compaction request fails with an API error (a rejection, an overload, a timeout or a connection error), or when the response holds no compaction block.

Arguments:

  • messages - The request to summarize, left unmodified.
  • tool_specs - The request’s tool specifications.
  • system_prompt - The request’s plain string system prompt.
  • system_prompt_content - The request’s structured system prompt.
  • instructions - Summarization instructions replacing the API’s own prompt.
  • cancel_signal - Event that aborts the base implementation’s call.
  • agent_metadata - Stable identity of the invoking agent, passed to the base implementation.
  • **kwargs - Additional keyword arguments for future extensibility.

Returns:

The summary and the usage of every call made for it.

@override
async def structured_output(
output_model: type[T],
prompt: Messages,
system_prompt: str | None = None,
**kwargs: Any) -> AsyncGenerator[dict[str, T | Any], None]

Defined in: src/strands/models/anthropic.py:1550

Get structured output from the model.

Arguments:

  • output_model - The output model to use for the agent.
  • prompt - The prompt messages to use for the agent.
  • system_prompt - System prompt to provide context to the model.
  • **kwargs - Additional keyword arguments for future extensibility.

Yields:

Model events with the last being the structured output.