diff --git a/README.md b/README.md
index 3a445a5..8a5033a 100644
--- a/README.md
+++ b/README.md
@@ -96,9 +96,15 @@ Follow our [Platform quickstart](https://docs.minitap.ai/mobile-use-sdk/platform
 
     Available models: `MiniMax-M2.7` (200K context, recommended), `MiniMax-M2.7-highspeed` (200K context, faster inference).
 
-    > [!NOTE]
-    > If you want to use Anthropic Claude, set `ANTHROPIC_API_KEY` in your `.env`.
-    >
+    **Using Anthropic Claude:**
+
+    [Anthropic](https://www.anthropic.com/) provides Claude — a capable, safety-focused model family. To use it:
+
+    1. Set `ANTHROPIC_API_KEY` in your `.env`
+    2. Copy the contents of the `anthropic` preset object from `llm-config.defaults.jsonc` into your `llm-config.override.jsonc`, or set `"provider": "anthropic"` on individual agent nodes.
+
+    Available models: `claude-sonnet-4-6` (200K context, recommended), `claude-haiku-4-5-20251001` (200K context, faster/cheaper).
+
     > [!NOTE]
     > If you want to use Google Vertex AI, you must either:
     >
diff --git a/llm-config.defaults.jsonc b/llm-config.defaults.jsonc
index 2f6987e..442aedf 100644
--- a/llm-config.defaults.jsonc
+++ b/llm-config.defaults.jsonc
@@ -71,6 +71,68 @@
       // }
     }
   },
+  // Anthropic preset – uses Claude Sonnet 4.6 (capable, 200k context) with Haiku 4.5 as a
+  // fast/cheap fallback. To use it, copy the contents of this object into llm-config.override.jsonc.
+  "anthropic": {
+    "planner": {
+      "provider": "anthropic",
+      "model": "claude-sonnet-4-6",
+      "fallback": {
+        "provider": "anthropic",
+        "model": "claude-haiku-4-5-20251001"
+      }
+    },
+    "orchestrator": {
+      "provider": "anthropic",
+      "model": "claude-sonnet-4-6",
+      "fallback": {
+        "provider": "anthropic",
+        "model": "claude-haiku-4-5-20251001"
+      }
+    },
+    "cortex": {
+      "provider": "anthropic",
+      "model": "claude-sonnet-4-6",
+      "fallback": {
+        "provider": "anthropic",
+        "model": "claude-haiku-4-5-20251001"
+      }
+    },
+    "executor": {
+      "provider": "anthropic",
+      "model": "claude-haiku-4-5-20251001",
+      "fallback": {
+        "provider": "anthropic",
+        "model": "claude-sonnet-4-6"
+      }
+    },
+    "contextor": {
+      "provider": "anthropic",
+      "model": "claude-haiku-4-5-20251001",
+      "fallback": {
+        "provider": "anthropic",
+        "model": "claude-sonnet-4-6"
+      }
+    },
+    "utils": {
+      "hopper": {
+        "provider": "anthropic",
+        "model": "claude-sonnet-4-6",
+        "fallback": {
+          "provider": "anthropic",
+          "model": "claude-haiku-4-5-20251001"
+        }
+      },
+      "outputter": {
+        "provider": "anthropic",
+        "model": "claude-haiku-4-5-20251001",
+        "fallback": {
+          "provider": "anthropic",
+          "model": "claude-sonnet-4-6"
+        }
+      }
+    }
+  },
   // MiniMax preset – uses MiniMax M2.7 and M2.7-highspeed (200k context).
   // To use it, copy the contents of this object into llm-config.override.jsonc.
   "minimax": {
diff --git a/minitap/mobile_use/services/llm.py b/minitap/mobile_use/services/llm.py
index 735f6a8..900bf60 100644
--- a/minitap/mobile_use/services/llm.py
+++ b/minitap/mobile_use/services/llm.py
@@ -145,6 +145,7 @@ def get_anthropic_llm(
         model_name=model_name,
         api_key=settings.ANTHROPIC_API_KEY,
         temperature=temperature,
+        max_tokens=8192,  # type: ignore[call-arg]  # field aliased to max_tokens_to_sample
         max_retries=2,
         timeout=None,
         stop=None,