support claude 4 sonnet (#235)

Signed-off-by: ChengZi <[email protected]>

Author: zc277584121

README.md

@@ -10,7 +10,7 @@
 
 ---
 
-DeepSearcher combines cutting-edge LLMs (OpenAI o3, Qwen3, DeepSeek, Grok 3, Claude 3.7 Sonnet, Llama 4, QwQ, etc.) and Vector Databases (Milvus, Zilliz Cloud etc.) to perform search, evaluation, and reasoning based on private data, providing highly accurate answer and comprehensive report. This project is suitable for enterprise knowledge management, intelligent Q&A systems, and information retrieval scenarios.
+DeepSearcher combines cutting-edge LLMs (OpenAI o3, Qwen3, DeepSeek, Grok 3, Claude 4 Sonnet, Llama 4, QwQ, etc.) and Vector Databases (Milvus, Zilliz Cloud etc.) to perform search, evaluation, and reasoning based on private data, providing highly accurate answer and comprehensive report. This project is suitable for enterprise knowledge management, intelligent Q&A systems, and information retrieval scenarios.
 
 ![Architecture](./assets/pic/deep-searcher-arch.png)
 
@@ -154,7 +154,7 @@ result = query("Write a report about xxx.") # Your question here
 <details>
   <summary>Example (Claude)</summary>
     <p> Make sure you have prepared your ANTHROPIC API KEY as an env variable <code>ANTHROPIC_API_KEY</code>.</p>
-    <pre><code>config.set_provider_config("llm", "Anthropic", {"model": "claude-3-7-sonnet-latest"})</code></pre>
+    <pre><code>config.set_provider_config("llm", "Anthropic", {"model": "claude-sonnet-4-0"})</code></pre>
     <p> More details about Anthropic Claude: https://docs.anthropic.com/en/home </p>
 </details>
 
@@ -485,7 +485,7 @@ Click on the button "Try it out", it allows you to fill the parameters and direc
 **Q1**: Why I failed to parse LLM output format / How to select the LLM?
 
 
-**A1**: Small LLMs struggle to follow the prompt to generate a desired response, which usually cause the format parsing problem. A better practice is to use large reasoning models e.g. deepseek-r1 671b, OpenAI o-series, Claude 3.7 sonnet, etc. as your LLM. 
+**A1**: Small LLMs struggle to follow the prompt to generate a desired response, which usually cause the format parsing problem. A better practice is to use large reasoning models e.g. deepseek-r1 671b, OpenAI o-series, Claude 4 sonnet, etc. as your LLM. 
 
 ---
 

deepsearcher/llm/anthropic_llm.py

@@ -17,12 +17,12 @@ class Anthropic(BaseLLM):
         client: The Anthropic client instance.
     """
 
-    def __init__(self, model: str = "claude-3-7-sonnet-latest", max_tokens: int = 8192, **kwargs):
+    def __init__(self, model: str = "claude-sonnet-4-0", max_tokens: int = 8192, **kwargs):
         """
         Initialize an Anthropic language model client.
 
         Args:
-            model (str, optional): The model identifier to use. Defaults to "claude-3-7-sonnet-latest".
+            model (str, optional): The model identifier to use. Defaults to "claude-sonnet-4-0".
             max_tokens (int, optional): The maximum number of tokens to generate. Defaults to 8192.
             **kwargs: Additional keyword arguments to pass to the Anthropic client.
                 - api_key: Anthropic API key. If not provided, uses ANTHROPIC_API_KEY environment variable.

docs/configuration/llm.md

@@ -14,7 +14,7 @@ config.set_provider_config("llm", "(LLMName)", "(Arguments dict)")
 |----------|-------------|------------|
 | **OpenAI** | OpenAI's API for GPT models | o1-mini, GPT-4 |
 | **DeepSeek** | DeepSeek AI offering | deepseek-reasoner, coder |
-| **Anthropic** | Anthropic's Claude models | claude-3-opus, claude-3-sonnet |
+| **Anthropic** | Anthropic's Claude models | claude-sonnet-4-0 |
 | **Gemini** | Google's Gemini models | gemini-1.5-pro, gemini-2.0-flash |
 | **XAI** | X.AI's Grok models | grok-2-latest |
 | **Ollama** | Local LLM deployment | llama3, qwq, etc. |
@@ -81,7 +81,7 @@ config.set_provider_config("llm", "DeepSeek", {"model": "deepseek-reasoner"})
 ??? example "Claude"
 
     ```python
-    config.set_provider_config("llm", "Anthropic", {"model": "claude-3-7-sonnet-latest"})
+    config.set_provider_config("llm", "Anthropic", {"model": "claude-sonnet-4-0"})
     ```
     *Requires `ANTHROPIC_API_KEY` environment variable*
 

docs/index.md

@@ -21,7 +21,7 @@
 
 ## ✨ Overview
 
-DeepSearcher combines cutting-edge LLMs (OpenAI o1, o3-mini, DeepSeek, Grok 3, Claude 3.7 Sonnet, Llama 4, QwQ, etc.) and Vector Databases (Milvus, Zilliz Cloud etc.) to perform search, evaluation, and reasoning based on private data, providing highly accurate answers and comprehensive reports.
+DeepSearcher combines cutting-edge LLMs (OpenAI o1, o3-mini, DeepSeek, Grok 3, Claude 4 Sonnet, Llama 4, QwQ, etc.) and Vector Databases (Milvus, Zilliz Cloud etc.) to perform search, evaluation, and reasoning based on private data, providing highly accurate answers and comprehensive reports.
 
 > **Perfect for:** Enterprise knowledge management, intelligent Q&A systems, and information retrieval scenarios.