Fix AnthropicModel.count_tokens with native tools

[kazmer97]

• Closes Anthropic count_tokens fails when native CodeExecutionTool is enabled #5702

Summary

• Omit Anthropic native/server tools from count_tokens requests by clearing native_tools on the count-only request parameters.
• Preserve native tools for normal message creation.
• Add a regression test covering CodeExecutionTool and WebFetchTool during preflight token counting.

Validation

• uv run pytest tests/models/test_anthropic.py::test_anthropic_count_tokens_with_mock tests/models/test_anthropic.py::test_anthropic_count_tokens_with_no_messages tests/models/test_anthropic.py::test_anthropic_count_tokens_omits_native_tools tests/models/test_anthropic.py::test_anthropic_code_execution_tool_version_auto tests/models/test_anthropic.py::test_anthropic_code_execution_tool_version_setting
• uv run ruff check pydantic_ai_slim/pydantic_ai/models/anthropic.py tests/models/test_anthropic.py
• uv run ruff format --check pydantic_ai_slim/pydantic_ai/models/anthropic.py tests/models/test_anthropic.py
• PYRIGHT_PYTHON_IGNORE_WARNINGS=1 uv run pyright pydantic_ai_slim/pydantic_ai/models/anthropic.py

Checklist

• Any AI generated code has been reviewed line-by-line by the human PR author, who stands by it.
• No breaking changes in accordance with the version policy.
• PR title is fit for the release changelog.

[github-actions]

Thanks for your interest in this issue! However, there are already open PRs addressing issue #5702: #5703.

To avoid duplicate efforts, this PR has been closed. If you'd like to contribute, you can review the existing PRs or share your thoughts on issue #5702.

If you believe the existing PRs are inactive, please comment on the issue and a maintainer can reassess.

[dsfaccini]

sorry about this @kazmer97 , sometimes issues/PRs get buried under spam, feel free to ping us directly on slack when this happens so we can act more quickly

let me do a review of the PR and get back to you, I may push updates if they're needed