Remove deprecated Claude 3.5 models (#4590)

crates/goose-bench/README.md

@@ -61,7 +61,7 @@ Benchmark configuration is provided through a JSON file. Here's a sample configu
     },
     {
       "provider": "databricks",
-      "name": "claude-3-5-sonnet",
+      "name": "claude-sonnet-4",
       "parallel_safe": true,
       "tool_shim": null
     },
@@ -232,7 +232,7 @@ Example of creating a config to re-run failed evaluations:
   "models": [
     {
       "provider": "databricks",
-      "name": "claude-3-5-sonnet",
+      "name": "claude-sonnet-4",
       "parallel_safe": false
     }
   ],

crates/goose-bench/src/bench_config.rs

@@ -51,7 +51,7 @@ impl Default for BenchRunConfig {
                 },
                 BenchModel {
                     provider: "databricks".to_string(),
-                    name: "goose-claude-3-5-sonnet".to_string(),
+                    name: "goose-claude-4-sonnet".to_string(),
                     parallel_safe: true,
                     tool_shim: None,
                 },