from __future__ import annotations as _annotations from typing import Literal, TypeAlias from ..native_tools import ( AdvisorTool, CodeExecutionTool, MCPServerTool, MemoryTool, WebFetchTool, WebSearchTool, ) from ..native_tools._tool_search import ToolSearchTool from ..settings import ThinkingEffort, ThinkingLevel from . import ModelProfile _ANTHROPIC_BASE_BUILTINS = frozenset({WebSearchTool, CodeExecutionTool, WebFetchTool, MemoryTool, MCPServerTool}) """Native tool types Anthropic generally supports across the model line. Mirrors `AnthropicModel.supported_native_tools()` minus `ToolSearchTool`, which is gated per-model in the profile below.""" ANTHROPIC_SAMPLING_PARAMS = ('temperature', 'top_p', 'top_k') """The unified sampling settings gated by `anthropic_disallows_sampling_settings`. Models whose profile sets that flag reject these settings outright with a 400, so every provider serving them has to drop and warn rather than forward. It lives here rather than beside either model because `models/bedrock.py` cannot import from `models/anthropic.py` without pulling the `anthropic` SDK into the `bedrock` extra. """ AnthropicCodeExecutionToolVersion: TypeAlias = Literal['20250825', '20260120'] """Concrete Anthropic code execution tool version to send for `CodeExecutionTool`.""" _ANTHROPIC_CODE_EXECUTION_20260120_MODEL_PREFIXES = ( 'claude-fable-5', 'claude-mythos-5', 'claude-opus-4-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-4-5', 'claude-sonnet-4-6', 'claude-sonnet-5', ) class AnthropicModelProfile(ModelProfile, total=False): """Profile for models used with `AnthropicModel`. ALL FIELDS MUST BE `anthropic_` PREFIXED SO YOU CAN MERGE THEM WITH OTHER MODELS. """ anthropic_supports_fast_speed: bool """Whether the model supports fast inference speed (`anthropic_speed='fast'`). Default: `False`. Currently Claude Opus 4.6, 4.7, 4.8, and 5 support fast mode. See the Anthropic docs for the latest list. """ anthropic_supports_adaptive_thinking: bool """Whether the model supports adaptive thinking (Sonnet 4.6+, Opus 4.6+). Default: `False`. When True, unified `thinking` translates to `{'type': 'adaptive'}`. When False, it translates to `{'type': 'enabled', 'budget_tokens': N}`. Because adaptive thinking — unlike extended thinking — is compatible with a forced `tool_choice`, this also decides whether unified `thinking` blocks tool forcing and switches Tool Output to Native or Prompted Output. """ anthropic_supports_effort: bool """Whether the model supports the `effort` parameter in `output_config` (Opus 4.5+, Sonnet 4.6+). Default: `False`. When True and the unified thinking level is a string (e.g. 'high'), it is also mapped to `output_config.effort`. """ anthropic_supports_dynamic_filtering: bool """Whether the model supports Anthropic-managed dynamic filtering for web search/fetch. Default: `False`. When enabled, Pydantic AI selects the `web_search_20260209` / `web_fetch_20260209` tool versions, which let Claude filter web results via code execution before they enter context. """ anthropic_supports_xhigh_effort: bool """Whether the model supports the `xhigh` effort value in `output_config`. Default: `False`. Claude Opus 4.7, 4.8, and 5 accept `xhigh`; older Anthropic models should use `max` instead. """ anthropic_disallows_budget_thinking: bool """Whether the model rejects budget-based thinking settings. Default: `False`. Claude Opus 4.7, 4.8, and 5 require adaptive thinking and return a 400 for `{'type': 'enabled', 'budget_tokens': ...}`. """ anthropic_disallows_sampling_settings: bool """Whether the model rejects sampling settings like `temperature` and `top_p`. Default: `False`. Claude Opus 4.7, 4.8, and 5 require these settings to be omitted from request payloads. """ anthropic_disallows_top_effort_when_thinking_disabled: bool """Whether the model rejects `xhigh`/`max` effort while thinking is explicitly disabled. Default: `False`. Claude Opus 5 caps effort at `high` when `anthropic_thinking={'type': 'disabled'}` and returns a 400 for `xhigh` or `max`; Claude Opus 4.8 accepts the same combination. """ anthropic_default_code_execution_tool_version: AnthropicCodeExecutionToolVersion """The Anthropic code execution tool version used when `anthropic_code_execution_tool_version='auto'`. Default: `'20250825'`.""" anthropic_supported_code_execution_tool_versions: tuple[AnthropicCodeExecutionToolVersion, ...] """The Anthropic code execution tool versions supported by the model. Default: `('20250825',)`.""" anthropic_supports_task_budgets: bool """Whether the model supports `output_config.task_budget`. Default: `False`. Anthropic currently documents task budgets as a Claude Opus 4.7 / 4.8 / 5 beta feature. """ anthropic_supports_forced_tool_choice: bool """Whether the model accepts a forced `tool_choice` (`{'type': 'any'}` or `{'type': 'tool'}`). Most Anthropic models only reject forcing alongside extended thinking; Claude Fable 5.1 and Claude Mythos 5.1 reject it unconditionally with a 400. When False, a resolved `required` tool choice falls back to `auto` (filtering tools to the requested set), and an explicit `tool_choice='required'` (or an explicit list of tools) raises a `UserError`. """ anthropic_binds_thinking_blocks: bool """Whether the model binds each thinking block to the conversation prefix that produced it. Default: `False`. Claude Fable 5.1 rejects a replayed thinking block once the `system` prompt text changes or a non-deferred tool joins the `tools` array — both of which Pydantic AI causes by design, through dynamic `@agent.instructions` and conditional toolsets. When True, Pydantic AI preserves the account's default behavior on the first request; if Anthropic rejects a stale block, it retries once with `thinking.block_binding.prefix_mismatch_behavior='drop_block'` and warns after the retry succeeds. """ ANTHROPIC_THINKING_BUDGET_MAP: dict[ThinkingLevel, int] = { True: 10000, 'minimal': 1024, 'low': 2048, 'medium': 10000, 'high': 16384, 'xhigh': 32768, } """Maps unified thinking values to Anthropic budget_tokens for non-adaptive models.""" AnthropicEffort: TypeAlias = Literal['low', 'medium', 'high', 'xhigh', 'max'] """Effort values Anthropic accepts at `output_config.effort`.""" ANTHROPIC_THINKING_EFFORT_MAP: dict[ThinkingEffort, AnthropicEffort] = { 'minimal': 'low', 'low': 'low', 'medium': 'medium', 'high': 'high', 'xhigh': 'max', } """Maps unified thinking effort levels to Anthropic `output_config.effort`. `xhigh` maps to `'max'` by default; callers that target a model with `anthropic_supports_xhigh_effort` should pass `supports_xhigh=True` to [`resolve_anthropic_effort`][pydantic_ai.profiles.anthropic.resolve_anthropic_effort] to preserve `xhigh` instead of downshifting. """ def resolve_anthropic_effort(level: ThinkingEffort, *, supports_xhigh: bool) -> AnthropicEffort: """Resolve a unified thinking effort level to the Anthropic `output_config.effort` value. Shared between the direct Anthropic path and any provider that translates to the Anthropic `output_config` wire shape (e.g. Bedrock Converse for Anthropic models). Keeps `ANTHROPIC_THINKING_EFFORT_MAP` as the single source of truth for the base mapping, while letting the `xhigh` passthrough decision live in one place. """ if level == 'xhigh' and supports_xhigh: return 'xhigh' return ANTHROPIC_THINKING_EFFORT_MAP[level] def anthropic_model_profile(model_name: str) -> ModelProfile | None: """Get the model profile for an Anthropic model.""" models_that_support_json_schema_output = ( 'claude-fable-5', 'claude-mythos-5', 'claude-haiku-4-5', 'claude-sonnet-4-5', 'claude-sonnet-4-6', 'claude-opus-4-1', 'claude-opus-4-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-5', ) """These models support both structured outputs and strict tool calling.""" # TODO update when new models are released that support structured outputs # https://docs.claude.com/en/docs/build-with-claude/structured-outputs#example-usage supports_json_schema_output = model_name.startswith(models_that_support_json_schema_output) anthropic_supports_fast_speed = model_name.startswith( ('claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5') ) # Sonnet 4.6+ and Opus 4.6+ support adaptive thinking; older models use budget-based supports_adaptive = model_name.startswith( ( 'claude-fable-5', 'claude-mythos-5', 'claude-sonnet-4-6', 'claude-sonnet-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', ) ) # Opus 4.5+ and Sonnet 4.6+ support the effort parameter in output_config supports_effort = model_name.startswith( ( 'claude-fable-5', 'claude-mythos-5', 'claude-opus-4-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-4-6', 'claude-sonnet-5', ) ) supports_xhigh_effort = model_name.startswith( ('claude-fable-5', 'claude-mythos-5', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-5') ) disallows_budget_thinking = model_name.startswith( ('claude-fable-5', 'claude-mythos-5', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-5') ) disallows_sampling_settings = model_name.startswith( ('claude-fable-5', 'claude-mythos-5', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-5') ) # Opus 5 caps effort at `high` while thinking is disabled; Opus 4.8 and earlier accept every level. disallows_top_effort_when_thinking_disabled = model_name.startswith('claude-opus-5') default_code_execution_tool_version, supported_code_execution_tool_versions = _code_execution_tool_versions( model_name ) supports_task_budgets = model_name.startswith( ('claude-fable-5', 'claude-mythos-5', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-5') ) # The 5.1 generation rejects a forced `tool_choice` (`any`/`tool`) outright, unlike other # Anthropic models which only reject forcing alongside extended thinking. Anthropic's # forcing-tool-use table names Claude Fable 5.1 and Claude Mythos 5.1 and nothing else, and # `claude-fable-5` accepts both forcing shapes live (200 on `any` and `tool`, GA and beta # endpoints), so Fable 5, Mythos 5, and Mythos Preview no longer belong here. supports_forced_tool_choice = not model_name.startswith(('claude-fable-5-1', 'claude-mythos-5-1')) # Claude Fable 5.1 alone binds thinking blocks to the conversation prefix: Claude Fable 5, # Opus 5, and Sonnet 5 all return 200 for a replayed block under an explicit # `prefix_mismatch_behavior` of `'error'`, and Anthropic documents that Claude Mythos 5.1 # "doesn't run this check" — the one capability on which it is not Fable 5.1's mirror. binds_thinking_blocks = model_name.startswith('claude-fable-5-1') supports_dynamic_filtering = model_name.startswith( ( 'claude-fable-5', 'claude-mythos-5', 'claude-mythos-preview', 'claude-sonnet-4-6', 'claude-sonnet-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', ) ) # Native tool search requires the `tool_search_tool_bm25_20251119` / # `tool_search_tool_regex_20251119` API types, which post-date Claude 4.0. In # practice, Anthropic enables it for Sonnet 4.5+, Opus 4.5+, and Haiku 4.5+. supports_tool_search = model_name.startswith( ( 'claude-fable-5', 'claude-mythos-5', 'claude-sonnet-4-5', 'claude-sonnet-4-6', 'claude-sonnet-5', 'claude-opus-4-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-haiku-4-5', ) ) # The advisor tool's valid *executors* (the model driving generation) are Fable/Mythos 5, Opus # 4.6-4.8 and 5, Sonnet 4.6/5, and Haiku 4.5. The executor/advisor pairing itself is validated API-side. supports_advisor = model_name.startswith( ( 'claude-fable-5', 'claude-mythos-5', 'claude-opus-4-6', 'claude-opus-4-7', 'claude-opus-4-8', 'claude-opus-5', 'claude-sonnet-4-6', 'claude-sonnet-5', 'claude-haiku-4-5', ) ) supported_native_tools = _ANTHROPIC_BASE_BUILTINS if supports_tool_search: supported_native_tools = supported_native_tools | {ToolSearchTool} if supports_advisor: supported_native_tools = supported_native_tools | {AdvisorTool} profile = AnthropicModelProfile( thinking_tags=('', ''), supports_json_schema_output=supports_json_schema_output, anthropic_supports_fast_speed=anthropic_supports_fast_speed, supports_thinking=True, anthropic_supports_adaptive_thinking=supports_adaptive, anthropic_supports_effort=supports_effort, anthropic_supports_dynamic_filtering=supports_dynamic_filtering, anthropic_supports_xhigh_effort=supports_xhigh_effort, anthropic_disallows_budget_thinking=disallows_budget_thinking, anthropic_disallows_sampling_settings=disallows_sampling_settings, anthropic_disallows_top_effort_when_thinking_disabled=disallows_top_effort_when_thinking_disabled, anthropic_default_code_execution_tool_version=default_code_execution_tool_version, anthropic_supported_code_execution_tool_versions=supported_code_execution_tool_versions, anthropic_supports_task_budgets=supports_task_budgets, anthropic_supports_forced_tool_choice=supports_forced_tool_choice, anthropic_binds_thinking_blocks=binds_thinking_blocks, supported_native_tools=supported_native_tools, ) if supports_tool_search: profile['tool_deferral_mode'] = 'standalone' return profile def _code_execution_tool_versions( model_name: str, ) -> tuple[AnthropicCodeExecutionToolVersion, tuple[AnthropicCodeExecutionToolVersion, ...]]: versions: tuple[AnthropicCodeExecutionToolVersion, ...] = ('20250825',) default_version: AnthropicCodeExecutionToolVersion = '20250825' if model_name.startswith(_ANTHROPIC_CODE_EXECUTION_20260120_MODEL_PREFIXES): default_version = '20260120' versions = (*versions, default_version) return default_version, versions