diff --git a/.env.example b/.env.example
index b025de2..a7e6376 100644
--- a/.env.example
+++ b/.env.example
@@ -58,10 +58,10 @@ DEFAULT_THINKING_MODE_THINKDEEP=high
 #   - mini             (shorthand for o4-mini)
 #
 # Supported Google/Gemini models:
-#   - gemini-2.5-flash-preview-05-20   (1M context, fast, supports thinking)
-#   - gemini-2.5-pro-preview-06-05     (1M context, powerful, supports thinking)
-#   - flash                             (shorthand for gemini-2.5-flash-preview-05-20)
-#   - pro                               (shorthand for gemini-2.5-pro-preview-06-05)
+#   - gemini-2.5-flash   (1M context, fast, supports thinking)
+#   - gemini-2.5-pro     (1M context, powerful, supports thinking)
+#   - flash                             (shorthand for gemini-2.5-flash)
+#   - pro                               (shorthand for gemini-2.5-pro)
 #
 # Supported X.AI GROK models:
 #   - grok-3          (131K context, advanced reasoning)
diff --git a/conf/custom_models.json b/conf/custom_models.json
index fb69e0c..1466451 100644
--- a/conf/custom_models.json
+++ b/conf/custom_models.json
@@ -89,7 +89,7 @@
       "description": "Google's Gemini 2.5 Pro via OpenRouter with vision"
     },
     {
-      "model_name": "google/gemini-2.5-flash-preview-05-20",
+      "model_name": "google/gemini-2.5-flash",
       "aliases": ["flash","gemini-flash", "flash-openrouter", "flash-2.5"],
       "context_window": 1048576,
       "supports_extended_thinking": false,
diff --git a/config.py b/config.py
index a9aa5ab..bf39757 100644
--- a/config.py
+++ b/config.py
@@ -14,7 +14,7 @@ import os
 # These values are used in server responses and for tracking releases
 # IMPORTANT: This is the single source of truth for version and author info
 # Semantic versioning: MAJOR.MINOR.PATCH
-__version__ = "5.1.0"
+__version__ = "5.1.2"
 # Last update date in ISO format
 __updated__ = "2025-06-18"
 # Primary maintainer
@@ -60,8 +60,8 @@ MODEL_CAPABILITIES_DESC = {
     "grok3": "GROK-3 (131K context) - Advanced reasoning model from X.AI, excellent for complex analysis",
     "grokfast": "GROK-3 Fast (131K context) - Higher performance variant, faster processing but more expensive",
     # Full model names also supported (for explicit specification)
-    "gemini-2.5-flash-preview-05-20": "Ultra-fast (1M context) - Quick analysis, simple queries, rapid iterations",
-    "gemini-2.5-pro-preview-06-05": (
+    "gemini-2.5-flash": "Ultra-fast (1M context) - Quick analysis, simple queries, rapid iterations",
+    "gemini-2.5-pro": (
         "Deep reasoning + thinking mode (1M context) - Complex problems, architecture, deep analysis"
     ),
 }
@@ -69,8 +69,8 @@ MODEL_CAPABILITIES_DESC = {
 # OpenRouter/Custom API Fallback Behavior:
 # When only OpenRouter or Custom API is configured (no native API keys), these
 # model aliases automatically map to equivalent models through the proxy:
-# - "flash" → "google/gemini-2.5-flash-preview-05-20" (via OpenRouter)
-# - "pro" → "google/gemini-2.5-pro-preview-06-05" (via OpenRouter)
+# - "flash" → "google/gemini-2.5-flash" (via OpenRouter)
+# - "pro" → "google/gemini-2.5-pro" (via OpenRouter)
 # - "o3" → "openai/o3" (via OpenRouter)
 # - "o3mini" → "openai/o3-mini" (via OpenRouter)
 # - "o4-mini" → "openai/o4-mini" (via OpenRouter)
diff --git a/docs/configuration.md b/docs/configuration.md
index f9cada6..8107cc4 100644
--- a/docs/configuration.md
+++ b/docs/configuration.md
@@ -124,8 +124,8 @@ OPENROUTER_ALLOWED_MODELS=opus,sonnet,mistral
 - `mini` (shorthand for o4-mini)
 
 **Gemini Models:**
-- `gemini-2.5-flash-preview-05-20` (1M context, fast)
-- `gemini-2.5-pro-preview-06-05` (1M context, powerful)
+- `gemini-2.5-flash` (1M context, fast)
+- `gemini-2.5-pro` (1M context, powerful)
 - `flash` (shorthand for Flash model)
 - `pro` (shorthand for Pro model)
 
diff --git a/docs/custom_models.md b/docs/custom_models.md
index 4bcd32e..8094675 100644
--- a/docs/custom_models.md
+++ b/docs/custom_models.md
@@ -46,8 +46,8 @@ The server uses `conf/custom_models.json` to map convenient aliases to both Open
 | `haiku` | `anthropic/claude-3-haiku` |
 | `gpt4o`, `4o` | `openai/gpt-4o` |
 | `gpt4o-mini`, `4o-mini` | `openai/gpt-4o-mini` |
-| `pro`, `gemini` | `google/gemini-2.5-pro-preview-06-05` |
-| `flash` | `google/gemini-2.5-flash-preview-05-20` |
+| `pro`, `gemini` | `google/gemini-2.5-pro` |
+| `flash` | `google/gemini-2.5-flash` |
 | `mistral` | `mistral/mistral-large` |
 | `deepseek`, `coder` | `deepseek/deepseek-coder` |
 | `perplexity` | `perplexity/llama-3-sonar-large-32k-online` |
@@ -153,7 +153,7 @@ CUSTOM_MODEL_NAME=your-loaded-model
 # OpenRouter models:
 "Use opus for deep analysis"         # → anthropic/claude-3-opus
 "Use sonnet to review this code"     # → anthropic/claude-3-sonnet
-"Use pro via zen to analyze this"    # → google/gemini-2.5-pro-preview-06-05
+"Use pro via zen to analyze this"    # → google/gemini-2.5-pro
 "Use gpt4o via zen to analyze this"  # → openai/gpt-4o
 "Use mistral via zen to optimize"    # → mistral/mistral-large
 
diff --git a/docs/tools/listmodels.md b/docs/tools/listmodels.md
index 1805ae3..59bbb37 100644
--- a/docs/tools/listmodels.md
+++ b/docs/tools/listmodels.md
@@ -46,7 +46,7 @@ The tool displays:
 📋 Available Models by Provider
 
 🔹 Google (Gemini) - ✅ Configured
-  • pro (gemini-2.5-pro-preview-06-05) - 1M context, thinking modes
+  • pro (gemini-2.5-pro) - 1M context, thinking modes
   • flash (gemini-2.0-flash-experimental) - 1M context, ultra-fast
 
 🔹 OpenAI - ✅ Configured  
diff --git a/providers/gemini.py b/providers/gemini.py
index 07e002a..4530481 100644
--- a/providers/gemini.py
+++ b/providers/gemini.py
@@ -19,14 +19,14 @@ class GeminiModelProvider(ModelProvider):
 
     # Model configurations
     SUPPORTED_MODELS = {
-        "gemini-2.5-flash-preview-05-20": {
+        "gemini-2.5-flash": {
             "context_window": 1_048_576,  # 1M tokens
             "supports_extended_thinking": True,
             "max_thinking_tokens": 24576,  # Flash 2.5 thinking budget limit
             "supports_images": True,  # Vision capability
             "max_image_size_mb": 20.0,  # Conservative 20MB limit for reliability
         },
-        "gemini-2.5-pro-preview-06-05": {
+        "gemini-2.5-pro": {
             "context_window": 1_048_576,  # 1M tokens
             "supports_extended_thinking": True,
             "max_thinking_tokens": 32768,  # Pro 2.5 thinking budget limit
@@ -34,8 +34,8 @@ class GeminiModelProvider(ModelProvider):
             "max_image_size_mb": 32.0,  # Higher limit for Pro model
         },
         # Shorthands
-        "flash": "gemini-2.5-flash-preview-05-20",
-        "pro": "gemini-2.5-pro-preview-06-05",
+        "flash": "gemini-2.5-flash",
+        "pro": "gemini-2.5-pro",
     }
 
     # Thinking mode configurations - percentages of model's max_thinking_tokens
@@ -364,8 +364,8 @@ class GeminiModelProvider(ModelProvider):
         """Check if the model supports vision (image processing)."""
         # Gemini 2.5 models support vision
         vision_models = {
-            "gemini-2.5-flash-preview-05-20",
-            "gemini-2.5-pro-preview-06-05",
+            "gemini-2.5-flash",
+            "gemini-2.5-pro",
             "gemini-2.0-flash",
             "gemini-1.5-pro",
             "gemini-1.5-flash",
diff --git a/providers/registry.py b/providers/registry.py
index 801f15f..4050d1c 100644
--- a/providers/registry.py
+++ b/providers/registry.py
@@ -105,7 +105,7 @@ class ModelProviderRegistry:
         3. OPENROUTER - Catch-all for cloud models via unified API
 
         Args:
-            model_name: Name of the model (e.g., "gemini-2.5-flash-preview-05-20", "o3-mini")
+            model_name: Name of the model (e.g., "gemini-2.5-flash", "o3-mini")
 
         Returns:
             ModelProvider instance that supports this model
@@ -295,7 +295,7 @@ class ModelProviderRegistry:
                 return custom_models[0]
             else:
                 # Fallback to pro if nothing found
-                return "gemini-2.5-pro-preview-06-05"
+                return "gemini-2.5-pro"
 
         elif tool_category == ToolModelCategory.FAST_RESPONSE:
             # Prefer fast, cost-efficient models
@@ -325,7 +325,7 @@ class ModelProviderRegistry:
                 return custom_models[0]
             else:
                 # Default to flash
-                return "gemini-2.5-flash-preview-05-20"
+                return "gemini-2.5-flash"
 
         # BALANCED or no category specified - use existing balanced logic
         if openai_available and "o4-mini" in openai_models:
@@ -353,7 +353,7 @@ class ModelProviderRegistry:
                 # This might happen if all models are restricted
                 logging.warning("No models available due to restrictions")
             # Return a reasonable default for backward compatibility
-            return "gemini-2.5-flash-preview-05-20"
+            return "gemini-2.5-flash"
 
     @classmethod
     def _find_extended_thinking_model(cls) -> Optional[str]:
@@ -383,7 +383,7 @@ class ModelProviderRegistry:
             preferred_models = [
                 "anthropic/claude-3.5-sonnet",
                 "anthropic/claude-3-opus-20240229",
-                "google/gemini-2.5-pro-preview-06-05",
+                "google/gemini-2.5-pro",
                 "google/gemini-pro-1.5",
                 "meta-llama/llama-3.1-70b-instruct",
                 "mistralai/mixtral-8x7b-instruct",
diff --git a/simulator_tests/test_model_thinking_config.py b/simulator_tests/test_model_thinking_config.py
index 94105c8..903bc3d 100644
--- a/simulator_tests/test_model_thinking_config.py
+++ b/simulator_tests/test_model_thinking_config.py
@@ -30,7 +30,7 @@ class TestModelThinkingConfig(BaseSimulatorTest):
                 "chat",
                 {
                     "prompt": "What is 2 + 2? Please think carefully and explain.",
-                    "model": "pro",  # Should resolve to gemini-2.5-pro-preview-06-05
+                    "model": "pro",  # Should resolve to gemini-2.5-pro
                     "thinking_mode": "high",  # Should use thinking_config
                 },
             )
@@ -55,7 +55,7 @@ class TestModelThinkingConfig(BaseSimulatorTest):
                 "chat",
                 {
                     "prompt": "What is 3 + 3? Give a quick answer.",
-                    "model": "flash",  # Should resolve to gemini-2.5-flash-preview-05-20
+                    "model": "flash",  # Should resolve to gemini-2.5-flash
                     "thinking_mode": "high",  # Should be ignored for Flash model
                 },
             )
@@ -79,8 +79,8 @@ class TestModelThinkingConfig(BaseSimulatorTest):
         test_cases = [
             ("pro", "should work with Pro model"),
             ("flash", "should work with Flash model"),
-            ("gemini-2.5-pro-preview-06-05", "should work with full Pro model name"),
-            ("gemini-2.5-flash-preview-05-20", "should work with full Flash model name"),
+            ("gemini-2.5-pro", "should work with full Pro model name"),
+            ("gemini-2.5-flash", "should work with full Flash model name"),
         ]
 
         success_count = 0
diff --git a/simulator_tests/test_openrouter_models.py b/simulator_tests/test_openrouter_models.py
index 1cd421c..5a52efd 100644
--- a/simulator_tests/test_openrouter_models.py
+++ b/simulator_tests/test_openrouter_models.py
@@ -43,7 +43,7 @@ class OpenRouterModelsTest(BaseSimulatorTest):
             self.setup_test_files()
 
             # Test 1: Flash alias mapping to OpenRouter
-            self.logger.info("  1: Testing 'flash' alias (should map to google/gemini-2.5-flash-preview-05-20)")
+            self.logger.info("  1: Testing 'flash' alias (should map to google/gemini-2.5-flash)")
 
             response1, continuation_id = self.call_mcp_tool(
                 "chat",
@@ -63,7 +63,7 @@ class OpenRouterModelsTest(BaseSimulatorTest):
                 self.logger.info(f"  ✅ Got continuation_id: {continuation_id}")
 
             # Test 2: Pro alias mapping to OpenRouter
-            self.logger.info("  2: Testing 'pro' alias (should map to google/gemini-2.5-pro-preview-06-05)")
+            self.logger.info("  2: Testing 'pro' alias (should map to google/gemini-2.5-pro)")
 
             response2, _ = self.call_mcp_tool(
                 "chat",
diff --git a/tests/conftest.py b/tests/conftest.py
index 0387bcd..ee5fe0b 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -25,7 +25,7 @@ if "XAI_API_KEY" not in os.environ:
 
 # Set default model to a specific value for tests to avoid auto mode
 # This prevents all tests from failing due to missing model parameter
-os.environ["DEFAULT_MODEL"] = "gemini-2.5-flash-preview-05-20"
+os.environ["DEFAULT_MODEL"] = "gemini-2.5-flash"
 
 # Force reload of config module to pick up the env var
 import config  # noqa: E402
@@ -108,7 +108,7 @@ def mock_provider_availability(request, monkeypatch):
         if model_name in ["unavailable-model", "gpt-5-turbo", "o3"]:
             return None
         # For common test models, return a mock provider
-        if model_name in ["gemini-2.5-flash-preview-05-20", "gemini-2.5-pro-preview-06-05", "pro", "flash"]:
+        if model_name in ["gemini-2.5-flash", "gemini-2.5-pro", "pro", "flash"]:
             # Try to use the real provider first if it exists
             real_provider = original_get_provider(model_name)
             if real_provider:
diff --git a/tests/mock_helpers.py b/tests/mock_helpers.py
index 0ee4465..eb283b6 100644
--- a/tests/mock_helpers.py
+++ b/tests/mock_helpers.py
@@ -5,7 +5,7 @@ from unittest.mock import Mock
 from providers.base import ModelCapabilities, ProviderType, RangeTemperatureConstraint
 
 
-def create_mock_provider(model_name="gemini-2.5-flash-preview-05-20", context_window=1_048_576):
+def create_mock_provider(model_name="gemini-2.5-flash", context_window=1_048_576):
     """Create a properly configured mock provider."""
     mock_provider = Mock()
 
diff --git a/tests/test_alias_target_restrictions.py b/tests/test_alias_target_restrictions.py
index 7641087..7b182e6 100644
--- a/tests/test_alias_target_restrictions.py
+++ b/tests/test_alias_target_restrictions.py
@@ -39,9 +39,9 @@ class TestAliasTargetRestrictions:
 
         # Should include both aliases and their targets
         assert "flash" in all_known  # alias
-        assert "gemini-2.5-flash-preview-05-20" in all_known  # target of 'flash'
+        assert "gemini-2.5-flash" in all_known  # target of 'flash'
         assert "pro" in all_known  # alias
-        assert "gemini-2.5-pro-preview-06-05" in all_known  # target of 'pro'
+        assert "gemini-2.5-pro" in all_known  # target of 'pro'
 
     @patch.dict(os.environ, {"OPENAI_ALLOWED_MODELS": "o4-mini"})  # Allow target
     def test_restriction_policy_allows_alias_when_target_allowed(self):
@@ -80,7 +80,7 @@ class TestAliasTargetRestrictions:
         # Direct target should NOT be allowed
         assert not provider.validate_model_name("o4-mini")
 
-    @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash-preview-05-20"})  # Allow target
+    @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash"})  # Allow target
     def test_gemini_restriction_policy_allows_alias_when_target_allowed(self):
         """Test Gemini restriction policy allows alias when target is allowed."""
         # Clear cached restriction service
@@ -91,7 +91,7 @@ class TestAliasTargetRestrictions:
         provider = GeminiModelProvider(api_key="test-key")
 
         # Both target and alias should be allowed
-        assert provider.validate_model_name("gemini-2.5-flash-preview-05-20")
+        assert provider.validate_model_name("gemini-2.5-flash")
         assert provider.validate_model_name("flash")
 
     @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "flash"})  # Allow alias only
@@ -107,7 +107,7 @@ class TestAliasTargetRestrictions:
         # Only the alias should be allowed
         assert provider.validate_model_name("flash")
         # Direct target should NOT be allowed
-        assert not provider.validate_model_name("gemini-2.5-flash-preview-05-20")
+        assert not provider.validate_model_name("gemini-2.5-flash")
 
     def test_restriction_service_validation_includes_all_targets(self):
         """Test that restriction service validation knows about all aliases and targets."""
diff --git a/tests/test_auto_mode_comprehensive.py b/tests/test_auto_mode_comprehensive.py
index 3f1e7dd..a6f281f 100644
--- a/tests/test_auto_mode_comprehensive.py
+++ b/tests/test_auto_mode_comprehensive.py
@@ -80,9 +80,9 @@ class TestAutoModeComprehensive:
                     "OPENROUTER_API_KEY": None,
                 },
                 {
-                    "EXTENDED_REASONING": "gemini-2.5-pro-preview-06-05",  # Pro for deep thinking
-                    "FAST_RESPONSE": "gemini-2.5-flash-preview-05-20",  # Flash for speed
-                    "BALANCED": "gemini-2.5-flash-preview-05-20",  # Flash as balanced
+                    "EXTENDED_REASONING": "gemini-2.5-pro",  # Pro for deep thinking
+                    "FAST_RESPONSE": "gemini-2.5-flash",  # Flash for speed
+                    "BALANCED": "gemini-2.5-flash",  # Flash as balanced
                 },
             ),
             # Only OpenAI API available
@@ -313,8 +313,8 @@ class TestAutoModeComprehensive:
             # Should include Gemini models
             assert "flash" in available_models
             assert "pro" in available_models
-            assert "gemini-2.5-flash-preview-05-20" in available_models
-            assert "gemini-2.5-pro-preview-06-05" in available_models
+            assert "gemini-2.5-flash" in available_models
+            assert "gemini-2.5-pro" in available_models
 
             # Should also include other models (users might have OpenRouter configured)
             # The schema should show all options; validation happens at runtime
@@ -476,8 +476,8 @@ class TestAutoModeComprehensive:
             assert "o3-mini" not in available_models
 
             # Should still include all Gemini models (no restrictions)
-            assert "gemini-2.5-flash-preview-05-20" in available_models
-            assert "gemini-2.5-pro-preview-06-05" in available_models
+            assert "gemini-2.5-flash" in available_models
+            assert "gemini-2.5-pro" in available_models
 
     def test_openrouter_fallback_when_no_native_apis(self):
         """Test that OpenRouter provides fallback models when no native APIs are available."""
@@ -511,8 +511,8 @@ class TestAutoModeComprehensive:
             # Mock OpenRouter registry to return known models
             mock_registry = MagicMock()
             mock_registry.list_models.return_value = [
-                "google/gemini-2.5-flash-preview-05-20",
-                "google/gemini-2.5-pro-preview-06-05",
+                "google/gemini-2.5-flash",
+                "google/gemini-2.5-pro",
                 "openai/o3",
                 "openai/o4-mini",
                 "anthropic/claude-3-opus",
@@ -564,11 +564,11 @@ class TestAutoModeComprehensive:
             mock_provider = MagicMock()
             mock_response = MagicMock()
             mock_response.content = "test response"
-            mock_response.model_name = "gemini-2.5-flash-preview-05-20"  # The resolved name
+            mock_response.model_name = "gemini-2.5-flash"  # The resolved name
             mock_response.usage = {"input_tokens": 10, "output_tokens": 5}
             # Mock _resolve_model_name to simulate alias resolution
             mock_provider._resolve_model_name = lambda alias: (
-                "gemini-2.5-flash-preview-05-20" if alias == "flash" else alias
+                "gemini-2.5-flash" if alias == "flash" else alias
             )
             mock_provider.generate_content.return_value = mock_response
 
diff --git a/tests/test_auto_mode_custom_provider_only.py b/tests/test_auto_mode_custom_provider_only.py
index 9b3abc7..5d03d4e 100644
--- a/tests/test_auto_mode_custom_provider_only.py
+++ b/tests/test_auto_mode_custom_provider_only.py
@@ -201,7 +201,7 @@ class TestAutoModeCustomProviderOnly:
 
                 # Should get a valid model name, not the hardcoded fallback
                 assert (
-                    fallback_model != "gemini-2.5-flash-preview-05-20"
+                    fallback_model != "gemini-2.5-flash"
                 ), "Should not fallback to hardcoded Gemini model when custom provider is available"
 
             except Exception as e:
diff --git a/tests/test_auto_mode_provider_selection.py b/tests/test_auto_mode_provider_selection.py
index 416d0d9..f610be4 100644
--- a/tests/test_auto_mode_provider_selection.py
+++ b/tests/test_auto_mode_provider_selection.py
@@ -59,9 +59,9 @@ class TestAutoModeProviderSelection:
             balanced = ModelProviderRegistry.get_preferred_fallback_model(ToolModelCategory.BALANCED)
 
             # Should select appropriate Gemini models
-            assert extended_reasoning in ["gemini-2.5-pro-preview-06-05", "pro"]
-            assert fast_response in ["gemini-2.5-flash-preview-05-20", "flash"]
-            assert balanced in ["gemini-2.5-flash-preview-05-20", "flash"]
+            assert extended_reasoning in ["gemini-2.5-pro", "pro"]
+            assert fast_response in ["gemini-2.5-flash", "flash"]
+            assert balanced in ["gemini-2.5-flash", "flash"]
 
         finally:
             # Restore original environment
@@ -229,8 +229,8 @@ class TestAutoModeProviderSelection:
             assert "o3-mini" not in available_models
 
             # Should include all Gemini models (no restrictions)
-            assert "gemini-2.5-flash-preview-05-20" in available_models
-            assert available_models["gemini-2.5-flash-preview-05-20"] == ProviderType.GOOGLE
+            assert "gemini-2.5-flash" in available_models
+            assert available_models["gemini-2.5-flash"] == ProviderType.GOOGLE
 
         finally:
             # Restore original environment
@@ -316,8 +316,8 @@ class TestAutoModeProviderSelection:
 
             # Test that providers resolve aliases correctly
             test_cases = [
-                ("flash", ProviderType.GOOGLE, "gemini-2.5-flash-preview-05-20"),
-                ("pro", ProviderType.GOOGLE, "gemini-2.5-pro-preview-06-05"),
+                ("flash", ProviderType.GOOGLE, "gemini-2.5-flash"),
+                ("pro", ProviderType.GOOGLE, "gemini-2.5-pro"),
                 ("mini", ProviderType.OPENAI, "o4-mini"),
                 ("o3mini", ProviderType.OPENAI, "o3-mini"),
                 ("grok", ProviderType.XAI, "grok-3"),
diff --git a/tests/test_buggy_behavior_prevention.py b/tests/test_buggy_behavior_prevention.py
index 4a4810e..e960f1f 100644
--- a/tests/test_buggy_behavior_prevention.py
+++ b/tests/test_buggy_behavior_prevention.py
@@ -85,16 +85,16 @@ class TestBuggyBehaviorPrevention:
 
         # Verify both aliases and targets are included
         assert "flash" in all_known  # alias
-        assert "gemini-2.5-flash-preview-05-20" in all_known  # target
+        assert "gemini-2.5-flash" in all_known  # target
         assert "pro" in all_known  # alias
-        assert "gemini-2.5-pro-preview-06-05" in all_known  # target
+        assert "gemini-2.5-pro" in all_known  # target
 
         # Simulate admin restricting to target model names
         service = ModelRestrictionService()
         service.restrictions = {
             ProviderType.GOOGLE: {
-                "gemini-2.5-flash-preview-05-20",  # Target name restriction
-                "gemini-2.5-pro-preview-06-05",  # Target name restriction
+                "gemini-2.5-flash",  # Target name restriction
+                "gemini-2.5-pro",  # Target name restriction
             }
         }
 
@@ -105,8 +105,8 @@ class TestBuggyBehaviorPrevention:
             # Should NOT warn about these valid target models
             all_warnings = [str(call) for call in mock_logger.warning.call_args_list]
             for warning in all_warnings:
-                assert "gemini-2.5-flash-preview-05-20" not in warning or "not a recognized" not in warning
-                assert "gemini-2.5-pro-preview-06-05" not in warning or "not a recognized" not in warning
+                assert "gemini-2.5-flash" not in warning or "not a recognized" not in warning
+                assert "gemini-2.5-pro" not in warning or "not a recognized" not in warning
 
     def test_old_bug_policy_bypass_prevention(self):
         """
@@ -206,7 +206,7 @@ class TestBuggyBehaviorPrevention:
         """
         providers_to_test = [
             (OpenAIModelProvider(api_key="test-key"), "mini", "o4-mini"),
-            (GeminiModelProvider(api_key="test-key"), "flash", "gemini-2.5-flash-preview-05-20"),
+            (GeminiModelProvider(api_key="test-key"), "flash", "gemini-2.5-flash"),
         ]
 
         for provider, alias, target in providers_to_test:
diff --git a/tests/test_claude_continuation.py b/tests/test_claude_continuation.py
index e3cf622..bca9413 100644
--- a/tests/test_claude_continuation.py
+++ b/tests/test_claude_continuation.py
@@ -59,7 +59,7 @@ class TestClaudeContinuationOffers:
         # If providers are not registered yet, tool might detect auto mode
         self.tool = ClaudeContinuationTool()
         # Set default model to avoid effective auto mode
-        self.tool.default_model = "gemini-2.5-flash-preview-05-20"
+        self.tool.default_model = "gemini-2.5-flash"
 
     @patch("utils.conversation_memory.get_storage")
     @patch.dict("os.environ", {"PYTEST_CURRENT_TEST": ""}, clear=False)
@@ -67,7 +67,7 @@ class TestClaudeContinuationOffers:
         """Test that new conversations offer Claude continuation opportunity"""
         # Create tool AFTER providers are registered (in conftest.py fixture)
         tool = ClaudeContinuationTool()
-        tool.default_model = "gemini-2.5-flash-preview-05-20"
+        tool.default_model = "gemini-2.5-flash"
 
         mock_client = Mock()
         mock_storage.return_value = mock_client
@@ -80,7 +80,7 @@ class TestClaudeContinuationOffers:
             mock_provider.generate_content.return_value = Mock(
                 content="Analysis complete.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -137,7 +137,7 @@ class TestClaudeContinuationOffers:
             mock_provider.generate_content.return_value = Mock(
                 content="Continued analysis.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -170,7 +170,7 @@ class TestClaudeContinuationOffers:
             mock_provider.generate_content.return_value = Mock(
                 content="Analysis complete. The code looks good.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -212,7 +212,7 @@ I'd be happy to examine the error handling patterns in more detail if that would
             mock_provider.generate_content.return_value = Mock(
                 content=content_with_followup,
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -257,7 +257,7 @@ I'd be happy to examine the error handling patterns in more detail if that would
             mock_provider.generate_content.return_value = Mock(
                 content="Continued analysis complete.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -313,7 +313,7 @@ I'd be happy to examine the error handling patterns in more detail if that would
             mock_provider.generate_content.return_value = Mock(
                 content="Final response.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -336,7 +336,7 @@ class TestContinuationIntegration:
     def setup_method(self):
         self.tool = ClaudeContinuationTool()
         # Set default model to avoid effective auto mode
-        self.tool.default_model = "gemini-2.5-flash-preview-05-20"
+        self.tool.default_model = "gemini-2.5-flash"
 
     @patch("utils.conversation_memory.get_storage")
     @patch.dict("os.environ", {"PYTEST_CURRENT_TEST": ""}, clear=False)
@@ -364,7 +364,7 @@ class TestContinuationIntegration:
             mock_provider.generate_content.return_value = Mock(
                 content="Analysis result",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -417,7 +417,7 @@ class TestContinuationIntegration:
             mock_provider.generate_content.return_value = Mock(
                 content="Structure analysis done.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -454,7 +454,7 @@ class TestContinuationIntegration:
             mock_provider.generate_content.return_value = Mock(
                 content="Performance analysis done.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
 
diff --git a/tests/test_collaboration.py b/tests/test_collaboration.py
index 186c91a..3c03e7b 100644
--- a/tests/test_collaboration.py
+++ b/tests/test_collaboration.py
@@ -41,7 +41,7 @@ class TestDynamicContextRequests:
         mock_provider.get_provider_type.return_value = Mock(value="google")
         mock_provider.supports_thinking_mode.return_value = False
         mock_provider.generate_content.return_value = Mock(
-            content=clarification_json, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=clarification_json, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
         mock_get_provider.return_value = mock_provider
 
@@ -86,7 +86,7 @@ class TestDynamicContextRequests:
         mock_provider.get_provider_type.return_value = Mock(value="google")
         mock_provider.supports_thinking_mode.return_value = False
         mock_provider.generate_content.return_value = Mock(
-            content=normal_response, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=normal_response, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
         mock_get_provider.return_value = mock_provider
 
@@ -110,7 +110,7 @@ class TestDynamicContextRequests:
         mock_provider.get_provider_type.return_value = Mock(value="google")
         mock_provider.supports_thinking_mode.return_value = False
         mock_provider.generate_content.return_value = Mock(
-            content=malformed_json, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=malformed_json, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
         mock_get_provider.return_value = mock_provider
 
@@ -150,7 +150,7 @@ class TestDynamicContextRequests:
         mock_provider.get_provider_type.return_value = Mock(value="google")
         mock_provider.supports_thinking_mode.return_value = False
         mock_provider.generate_content.return_value = Mock(
-            content=clarification_json, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=clarification_json, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
         mock_get_provider.return_value = mock_provider
 
@@ -274,7 +274,7 @@ class TestCollaborationWorkflow:
         mock_provider.get_provider_type.return_value = Mock(value="google")
         mock_provider.supports_thinking_mode.return_value = False
         mock_provider.generate_content.return_value = Mock(
-            content=clarification_json, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=clarification_json, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
         mock_get_provider.return_value = mock_provider
 
@@ -313,7 +313,7 @@ class TestCollaborationWorkflow:
         mock_provider.get_provider_type.return_value = Mock(value="google")
         mock_provider.supports_thinking_mode.return_value = False
         mock_provider.generate_content.return_value = Mock(
-            content=clarification_json, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=clarification_json, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
         mock_get_provider.return_value = mock_provider
 
@@ -340,7 +340,7 @@ class TestCollaborationWorkflow:
         """
 
         mock_provider.generate_content.return_value = Mock(
-            content=final_response, usage={}, model_name="gemini-2.5-flash-preview-05-20", metadata={}
+            content=final_response, usage={}, model_name="gemini-2.5-flash", metadata={}
         )
 
         result2 = await tool.execute(
diff --git a/tests/test_config.py b/tests/test_config.py
index e6ad23b..ba793f9 100644
--- a/tests/test_config.py
+++ b/tests/test_config.py
@@ -31,7 +31,7 @@ class TestConfig:
     def test_model_config(self):
         """Test model configuration"""
         # DEFAULT_MODEL is set in conftest.py for tests
-        assert DEFAULT_MODEL == "gemini-2.5-flash-preview-05-20"
+        assert DEFAULT_MODEL == "gemini-2.5-flash"
 
     def test_temperature_defaults(self):
         """Test temperature constants"""
diff --git a/tests/test_conversation_field_mapping.py b/tests/test_conversation_field_mapping.py
index 589a732..6450b2c 100644
--- a/tests/test_conversation_field_mapping.py
+++ b/tests/test_conversation_field_mapping.py
@@ -74,7 +74,7 @@ async def test_conversation_history_field_mapping():
                         mock_provider = MagicMock()
                         mock_provider.get_capabilities.return_value = ModelCapabilities(
                             provider=ProviderType.GOOGLE,
-                            model_name="gemini-2.5-flash-preview-05-20",
+                            model_name="gemini-2.5-flash",
                             friendly_name="Gemini",
                             context_window=200000,
                             supports_extended_thinking=True,
@@ -132,7 +132,7 @@ async def test_unknown_tool_defaults_to_prompt():
                 # Mock ModelContext to avoid calculation errors
                 with patch("utils.model_context.ModelContext") as mock_model_context_class:
                     mock_model_context = MagicMock()
-                    mock_model_context.model_name = "gemini-2.5-flash-preview-05-20"
+                    mock_model_context.model_name = "gemini-2.5-flash"
                     mock_model_context.calculate_token_allocation.return_value = MagicMock(
                         total_tokens=200000,
                         content_tokens=120000,
diff --git a/tests/test_conversation_history_bug.py b/tests/test_conversation_history_bug.py
index c5c6386..efe3036 100644
--- a/tests/test_conversation_history_bug.py
+++ b/tests/test_conversation_history_bug.py
@@ -95,7 +95,7 @@ class TestConversationHistoryBugFix:
                 return Mock(
                     content="Response with conversation context",
                     usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                    model_name="gemini-2.5-flash-preview-05-20",
+                    model_name="gemini-2.5-flash",
                     metadata={"finish_reason": "STOP"},
                 )
 
@@ -155,7 +155,7 @@ class TestConversationHistoryBugFix:
                 return Mock(
                     content="Response without history",
                     usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                    model_name="gemini-2.5-flash-preview-05-20",
+                    model_name="gemini-2.5-flash",
                     metadata={"finish_reason": "STOP"},
                 )
 
@@ -193,7 +193,7 @@ class TestConversationHistoryBugFix:
                 return Mock(
                     content="New conversation response",
                     usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                    model_name="gemini-2.5-flash-preview-05-20",
+                    model_name="gemini-2.5-flash",
                     metadata={"finish_reason": "STOP"},
                 )
 
@@ -277,7 +277,7 @@ class TestConversationHistoryBugFix:
                 return Mock(
                     content="Analysis of new files complete",
                     usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                    model_name="gemini-2.5-flash-preview-05-20",
+                    model_name="gemini-2.5-flash",
                     metadata={"finish_reason": "STOP"},
                 )
 
diff --git a/tests/test_cross_tool_continuation.py b/tests/test_cross_tool_continuation.py
index 4f52464..23b95f4 100644
--- a/tests/test_cross_tool_continuation.py
+++ b/tests/test_cross_tool_continuation.py
@@ -112,7 +112,7 @@ I'd be happy to review these security findings in detail if that would be helpfu
             mock_provider.generate_content.return_value = Mock(
                 content=content,
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -159,7 +159,7 @@ I'd be happy to review these security findings in detail if that would be helpfu
             mock_provider.generate_content.return_value = Mock(
                 content="Critical security vulnerability confirmed. The authentication function always returns true, bypassing all security checks.",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -284,7 +284,7 @@ I'd be happy to review these security findings in detail if that would be helpfu
             mock_provider.generate_content.return_value = Mock(
                 content="Security review of auth.py shows vulnerabilities",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
diff --git a/tests/test_intelligent_fallback.py b/tests/test_intelligent_fallback.py
index 6d7637e..e79f2a5 100644
--- a/tests/test_intelligent_fallback.py
+++ b/tests/test_intelligent_fallback.py
@@ -48,14 +48,14 @@ class TestIntelligentFallback:
 
     @patch.dict(os.environ, {"OPENAI_API_KEY": "", "GEMINI_API_KEY": "test-gemini-key"}, clear=False)
     def test_prefers_gemini_flash_when_openai_unavailable(self):
-        """Test that gemini-2.5-flash-preview-05-20 is used when only Gemini API key is available"""
+        """Test that gemini-2.5-flash is used when only Gemini API key is available"""
         # Register only Gemini provider for this test
         from providers.gemini import GeminiModelProvider
 
         ModelProviderRegistry.register_provider(ProviderType.GOOGLE, GeminiModelProvider)
 
         fallback_model = ModelProviderRegistry.get_preferred_fallback_model()
-        assert fallback_model == "gemini-2.5-flash-preview-05-20"
+        assert fallback_model == "gemini-2.5-flash"
 
     @patch.dict(os.environ, {"OPENAI_API_KEY": "sk-test-key", "GEMINI_API_KEY": "test-gemini-key"}, clear=False)
     def test_prefers_openai_when_both_available(self):
@@ -81,7 +81,7 @@ class TestIntelligentFallback:
         ModelProviderRegistry.register_provider(ProviderType.GOOGLE, GeminiModelProvider)
 
         fallback_model = ModelProviderRegistry.get_preferred_fallback_model()
-        assert fallback_model == "gemini-2.5-flash-preview-05-20"  # Default fallback
+        assert fallback_model == "gemini-2.5-flash"  # Default fallback
 
     def test_available_providers_with_keys(self):
         """Test the get_available_providers_with_keys method"""
@@ -186,14 +186,14 @@ class TestIntelligentFallback:
 
                 history, tokens = build_conversation_history(context, model_context=None)
 
-                # Should use gemini-2.5-flash-preview-05-20 when only Gemini is available
-                mock_context_class.assert_called_once_with("gemini-2.5-flash-preview-05-20")
+                # Should use gemini-2.5-flash when only Gemini is available
+                mock_context_class.assert_called_once_with("gemini-2.5-flash")
 
     def test_non_auto_mode_unchanged(self):
         """Test that non-auto mode behavior is unchanged"""
         from utils.conversation_memory import ThreadContext, build_conversation_history
 
-        with patch("config.IS_AUTO_MODE", False), patch("config.DEFAULT_MODEL", "gemini-2.5-pro-preview-06-05"):
+        with patch("config.IS_AUTO_MODE", False), patch("config.DEFAULT_MODEL", "gemini-2.5-pro"):
             from utils.conversation_memory import ConversationTurn
 
             context = ThreadContext(
@@ -219,7 +219,7 @@ class TestIntelligentFallback:
                 history, tokens = build_conversation_history(context, model_context=None)
 
                 # Should use the configured DEFAULT_MODEL, not the intelligent fallback
-                mock_context_class.assert_called_once_with("gemini-2.5-pro-preview-06-05")
+                mock_context_class.assert_called_once_with("gemini-2.5-pro")
 
 
 if __name__ == "__main__":
diff --git a/tests/test_large_prompt_handling.py b/tests/test_large_prompt_handling.py
index 14e249a..afd8c7f 100644
--- a/tests/test_large_prompt_handling.py
+++ b/tests/test_large_prompt_handling.py
@@ -76,7 +76,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="This is a test response",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -103,7 +103,7 @@ class TestLargePromptHandling:
             patch("utils.model_context.ModelContext") as mock_model_context_class,
         ):
 
-            mock_provider = create_mock_provider(model_name="gemini-2.5-flash-preview-05-20", context_window=1_048_576)
+            mock_provider = create_mock_provider(model_name="gemini-2.5-flash", context_window=1_048_576)
             mock_provider.generate_content.return_value.content = "Processed prompt from file"
             mock_get_provider.return_value = mock_provider
 
@@ -111,7 +111,7 @@ class TestLargePromptHandling:
             from utils.model_context import TokenAllocation
 
             mock_model_context = MagicMock()
-            mock_model_context.model_name = "gemini-2.5-flash-preview-05-20"
+            mock_model_context.model_name = "gemini-2.5-flash"
             mock_model_context.calculate_token_allocation.return_value = TokenAllocation(
                 total_tokens=1_048_576,
                 content_tokens=838_861,
@@ -293,7 +293,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="Success",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -336,7 +336,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="Response to the large prompt",
                 usage={"input_tokens": 12000, "output_tokens": 10, "total_tokens": 12010},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -368,7 +368,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="Success",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -390,7 +390,7 @@ class TestLargePromptHandling:
             patch("utils.model_context.ModelContext") as mock_model_context_class,
         ):
 
-            mock_provider = create_mock_provider(model_name="gemini-2.5-flash-preview-05-20", context_window=1_048_576)
+            mock_provider = create_mock_provider(model_name="gemini-2.5-flash", context_window=1_048_576)
             mock_provider.generate_content.return_value.content = "Success"
             mock_get_provider.return_value = mock_provider
 
@@ -398,7 +398,7 @@ class TestLargePromptHandling:
             from utils.model_context import TokenAllocation
 
             mock_model_context = MagicMock()
-            mock_model_context.model_name = "gemini-2.5-flash-preview-05-20"
+            mock_model_context.model_name = "gemini-2.5-flash"
             mock_model_context.calculate_token_allocation.return_value = TokenAllocation(
                 total_tokens=1_048_576,
                 content_tokens=838_861,
@@ -437,7 +437,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="Weather is sunny",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -502,7 +502,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="Hi there!",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
@@ -555,7 +555,7 @@ class TestLargePromptHandling:
             mock_provider.generate_content.return_value = MagicMock(
                 content="Continuing our conversation...",
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
             mock_get_provider.return_value = mock_provider
diff --git a/tests/test_listmodels.py b/tests/test_listmodels.py
index d28f5f8..212672c 100644
--- a/tests/test_listmodels.py
+++ b/tests/test_listmodels.py
@@ -65,8 +65,8 @@ class TestListModelsTool:
 
             # Check Gemini shows as configured
             assert "Google Gemini ✅" in content
-            assert "`flash` → `gemini-2.5-flash-preview-05-20`" in content
-            assert "`pro` → `gemini-2.5-pro-preview-06-05`" in content
+            assert "`flash` → `gemini-2.5-flash`" in content
+            assert "`pro` → `gemini-2.5-pro`" in content
             assert "1M context" in content
 
             # Check summary
diff --git a/tests/test_model_enumeration.py b/tests/test_model_enumeration.py
index 7046574..544cdf1 100644
--- a/tests/test_model_enumeration.py
+++ b/tests/test_model_enumeration.py
@@ -96,8 +96,8 @@ class TestModelEnumeration:
             "grok-3-fast",
             "grok3",
             "grokfast",  # X.AI models
-            "gemini-2.5-flash-preview-05-20",
-            "gemini-2.5-pro-preview-06-05",  # Full Gemini names
+            "gemini-2.5-flash",
+            "gemini-2.5-pro",  # Full Gemini names
         ]
 
         for model in native_models:
@@ -264,7 +264,7 @@ class TestModelEnumeration:
             ("flash", True),  # Native Gemini
             ("o3", True),  # Native OpenAI
             ("grok", True),  # Native X.AI
-            ("gemini-2.5-flash-preview-05-20", True),  # Full native name
+            ("gemini-2.5-flash", True),  # Full native name
             ("o4-mini-high", True),  # Native OpenAI variant
             ("grok-3-fast", True),  # Native X.AI variant
         ],
diff --git a/tests/test_model_restrictions.py b/tests/test_model_restrictions.py
index 75af28b..0df646a 100644
--- a/tests/test_model_restrictions.py
+++ b/tests/test_model_restrictions.py
@@ -22,8 +22,8 @@ class TestModelRestrictionService:
             # Should allow all models
             assert service.is_allowed(ProviderType.OPENAI, "o3")
             assert service.is_allowed(ProviderType.OPENAI, "o3-mini")
-            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro-preview-06-05")
-            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-flash-preview-05-20")
+            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro")
+            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-flash")
             assert service.is_allowed(ProviderType.OPENROUTER, "anthropic/claude-3-opus")
             assert service.is_allowed(ProviderType.OPENROUTER, "openai/o3")
 
@@ -43,7 +43,7 @@ class TestModelRestrictionService:
             assert not service.is_allowed(ProviderType.OPENAI, "o4-mini")
 
             # Google and OpenRouter should have no restrictions
-            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro-preview-06-05")
+            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro")
             assert service.is_allowed(ProviderType.OPENROUTER, "anthropic/claude-3-opus")
 
     def test_load_multiple_models_restriction(self):
@@ -59,7 +59,7 @@ class TestModelRestrictionService:
             # Check Google models
             assert service.is_allowed(ProviderType.GOOGLE, "flash")
             assert service.is_allowed(ProviderType.GOOGLE, "pro")
-            assert not service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro-preview-06-05")
+            assert not service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro")
 
     def test_case_insensitive_and_whitespace_handling(self):
         """Test that model names are case-insensitive and whitespace is trimmed."""
@@ -84,9 +84,9 @@ class TestModelRestrictionService:
 
             # Google should only allow flash (and its resolved name)
             assert service.is_allowed(ProviderType.GOOGLE, "flash")
-            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-flash-preview-05-20", "flash")
+            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-flash", "flash")
             assert not service.is_allowed(ProviderType.GOOGLE, "pro")
-            assert not service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro-preview-06-05", "pro")
+            assert not service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro", "pro")
 
     def test_filter_models(self):
         """Test filtering a list of models based on restrictions."""
@@ -124,8 +124,8 @@ class TestModelRestrictionService:
             assert not service.is_allowed(ProviderType.OPENAI, "o3")
 
             # Google should allow both models via shorthands
-            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-flash-preview-05-20", "flash")
-            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro-preview-06-05", "pro")
+            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-flash", "flash")
+            assert service.is_allowed(ProviderType.GOOGLE, "gemini-2.5-pro", "pro")
 
             # Also test that full names work when specified in restrictions
             assert service.is_allowed(ProviderType.OPENAI, "o3-mini", "o3mini")  # Even with shorthand
@@ -238,7 +238,7 @@ class TestProviderIntegration:
             provider.get_capabilities("o3")
         assert "not allowed by restriction policy" in str(exc_info.value)
 
-    @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash-preview-05-20,flash"})
+    @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash,flash"})
     def test_gemini_provider_respects_restrictions(self):
         """Test that Gemini provider respects restrictions."""
         # Clear any cached restriction service
@@ -250,11 +250,11 @@ class TestProviderIntegration:
 
         # Should validate allowed models (both shorthand and full name allowed)
         assert provider.validate_model_name("flash")
-        assert provider.validate_model_name("gemini-2.5-flash-preview-05-20")
+        assert provider.validate_model_name("gemini-2.5-flash")
 
         # Should not validate disallowed model
         assert not provider.validate_model_name("pro")
-        assert not provider.validate_model_name("gemini-2.5-pro-preview-06-05")
+        assert not provider.validate_model_name("gemini-2.5-pro")
 
         # get_capabilities should raise for disallowed model
         with pytest.raises(ValueError) as exc_info:
@@ -288,13 +288,13 @@ class TestProviderIntegration:
 
         # Should allow getting capabilities for "flash"
         capabilities = provider.get_capabilities("flash")
-        assert capabilities.model_name == "gemini-2.5-flash-preview-05-20"
+        assert capabilities.model_name == "gemini-2.5-flash"
 
         # Test the edge case: Try to use full model name when only alias is allowed
         # This should NOT be allowed - only the alias "flash" is in the restriction list
-        assert not provider.validate_model_name("gemini-2.5-flash-preview-05-20")
+        assert not provider.validate_model_name("gemini-2.5-flash")
 
-    @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash-preview-05-20"})
+    @patch.dict(os.environ, {"GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash"})
     def test_gemini_parameter_order_edge_case_full_name_only(self):
         """Test parameter order with only full name allowed, not alias.
 
@@ -310,7 +310,7 @@ class TestProviderIntegration:
         provider = GeminiModelProvider(api_key="test-key")
 
         # Should allow full name
-        assert provider.validate_model_name("gemini-2.5-flash-preview-05-20")
+        assert provider.validate_model_name("gemini-2.5-flash")
 
         # Should also allow alias that resolves to allowed full name
         # This works because is_allowed checks both resolved_name and original_name
@@ -318,7 +318,7 @@ class TestProviderIntegration:
 
         # Should not allow "pro" alias
         assert not provider.validate_model_name("pro")
-        assert not provider.validate_model_name("gemini-2.5-pro-preview-06-05")
+        assert not provider.validate_model_name("gemini-2.5-pro")
 
 
 class TestCustomProviderOpenRouterRestrictions:
@@ -469,8 +469,8 @@ class TestRegistryIntegration:
 
         mock_gemini = MagicMock()
         mock_gemini.SUPPORTED_MODELS = {
-            "gemini-2.5-pro-preview-06-05": {"context_window": 1048576},
-            "gemini-2.5-flash-preview-05-20": {"context_window": 1048576},
+            "gemini-2.5-pro": {"context_window": 1048576},
+            "gemini-2.5-flash": {"context_window": 1048576},
         }
         mock_gemini.get_provider_type.return_value = ProviderType.GOOGLE
 
@@ -493,8 +493,8 @@ class TestRegistryIntegration:
 
         mock_gemini.list_models = gemini_list_models
         mock_gemini.list_all_known_models.return_value = [
-            "gemini-2.5-pro-preview-06-05",
-            "gemini-2.5-flash-preview-05-20",
+            "gemini-2.5-pro",
+            "gemini-2.5-flash",
         ]
 
         def get_provider_side_effect(provider_type):
@@ -514,7 +514,7 @@ class TestRegistryIntegration:
         }
 
         with patch.dict(
-            os.environ, {"OPENAI_ALLOWED_MODELS": "o3-mini", "GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash-preview-05-20"}
+            os.environ, {"OPENAI_ALLOWED_MODELS": "o3-mini", "GOOGLE_ALLOWED_MODELS": "gemini-2.5-flash"}
         ):
             # Clear cached restriction service
             import utils.model_restrictions
@@ -526,8 +526,8 @@ class TestRegistryIntegration:
             # Should only include allowed models
             assert "o3-mini" in available
             assert "o3" not in available
-            assert "gemini-2.5-flash-preview-05-20" in available
-            assert "gemini-2.5-pro-preview-06-05" not in available
+            assert "gemini-2.5-flash" in available
+            assert "gemini-2.5-pro" not in available
 
 
 class TestShorthandRestrictions:
@@ -552,7 +552,7 @@ class TestShorthandRestrictions:
         gemini_provider = GeminiModelProvider(api_key="test-key")
         assert gemini_provider.validate_model_name("flash")  # Should work with shorthand
         # Same for Gemini - if you restrict to "flash", you can't use the full name
-        assert not gemini_provider.validate_model_name("gemini-2.5-flash-preview-05-20")  # Not allowed
+        assert not gemini_provider.validate_model_name("gemini-2.5-flash")  # Not allowed
         assert not gemini_provider.validate_model_name("pro")  # Not allowed
 
     @patch.dict(os.environ, {"OPENAI_ALLOWED_MODELS": "o3mini,mini,o4-mini"})
@@ -579,7 +579,7 @@ class TestShorthandRestrictions:
 
     @patch.dict(
         os.environ,
-        {"OPENAI_ALLOWED_MODELS": "mini,o4-mini", "GOOGLE_ALLOWED_MODELS": "flash,gemini-2.5-flash-preview-05-20"},
+        {"OPENAI_ALLOWED_MODELS": "mini,o4-mini", "GOOGLE_ALLOWED_MODELS": "flash,gemini-2.5-flash"},
     )
     def test_both_shorthand_and_full_name_allowed(self):
         """Test that we can allow both shorthand and full names."""
@@ -596,7 +596,7 @@ class TestShorthandRestrictions:
         # Gemini - both flash and full name are allowed
         gemini_provider = GeminiModelProvider(api_key="test-key")
         assert gemini_provider.validate_model_name("flash")
-        assert gemini_provider.validate_model_name("gemini-2.5-flash-preview-05-20")
+        assert gemini_provider.validate_model_name("gemini-2.5-flash")
 
 
 class TestAutoModeWithRestrictions:
@@ -688,7 +688,7 @@ class TestAutoModeWithRestrictions:
 
             # The fallback will depend on how get_available_models handles aliases
             # For now, we accept either behavior and document it
-            assert model in ["o4-mini", "gemini-2.5-flash-preview-05-20"]
+            assert model in ["o4-mini", "gemini-2.5-flash"]
         finally:
             # Restore original registry state
             registry = ModelProviderRegistry()
diff --git a/tests/test_old_behavior_simulation.py b/tests/test_old_behavior_simulation.py
index 1f6ebc5..d14c8ea 100644
--- a/tests/test_old_behavior_simulation.py
+++ b/tests/test_old_behavior_simulation.py
@@ -167,7 +167,7 @@ class TestOldBehaviorSimulation:
             (["mini", "o3mini"], ["mini", "o3mini", "o4-mini", "o3-mini"], "OpenAI"),
             (
                 ["flash", "pro"],
-                ["flash", "pro", "gemini-2.5-flash-preview-05-20", "gemini-2.5-pro-preview-06-05"],
+                ["flash", "pro", "gemini-2.5-flash", "gemini-2.5-pro"],
                 "Gemini",
             ),
         ]
diff --git a/tests/test_openrouter_provider.py b/tests/test_openrouter_provider.py
index 41d45a6..544c464 100644
--- a/tests/test_openrouter_provider.py
+++ b/tests/test_openrouter_provider.py
@@ -151,8 +151,8 @@ class TestOpenRouterAutoMode:
 
         mock_registry = Mock()
         mock_registry.list_models.return_value = [
-            "google/gemini-2.5-flash-preview-05-20",
-            "google/gemini-2.5-pro-preview-06-05",
+            "google/gemini-2.5-flash",
+            "google/gemini-2.5-pro",
             "openai/o3",
             "openai/o3-mini",
             "anthropic/claude-3-opus",
@@ -181,7 +181,7 @@ class TestOpenRouterAutoMode:
         os.environ.pop("OPENAI_API_KEY", None)
         os.environ["OPENROUTER_API_KEY"] = "test-openrouter-key"
         os.environ.pop("OPENROUTER_ALLOWED_MODELS", None)
-        os.environ["OPENROUTER_ALLOWED_MODELS"] = "anthropic/claude-3-opus,google/gemini-2.5-flash-preview-05-20"
+        os.environ["OPENROUTER_ALLOWED_MODELS"] = "anthropic/claude-3-opus,google/gemini-2.5-flash"
         os.environ["DEFAULT_MODEL"] = "auto"
 
         # Force reload to pick up new environment variable
@@ -191,8 +191,8 @@ class TestOpenRouterAutoMode:
 
         mock_registry = Mock()
         mock_registry.list_models.return_value = [
-            "google/gemini-2.5-flash-preview-05-20",
-            "google/gemini-2.5-pro-preview-06-05",
+            "google/gemini-2.5-flash",
+            "google/gemini-2.5-pro",
             "anthropic/claude-3-opus",
             "anthropic/claude-3-sonnet",
         ]
@@ -206,7 +206,7 @@ class TestOpenRouterAutoMode:
 
         assert len(available_models) > 0, "Should have some allowed models"
 
-        expected_allowed = {"google/gemini-2.5-flash-preview-05-20", "anthropic/claude-3-opus"}
+        expected_allowed = {"google/gemini-2.5-flash", "anthropic/claude-3-opus"}
 
         assert (
             set(available_models.keys()) == expected_allowed
diff --git a/tests/test_per_tool_model_defaults.py b/tests/test_per_tool_model_defaults.py
index c2af732..9354588 100644
--- a/tests/test_per_tool_model_defaults.py
+++ b/tests/test_per_tool_model_defaults.py
@@ -91,13 +91,13 @@ class TestModelSelection:
         with patch.object(ModelProviderRegistry, "get_available_models") as mock_get_available:
             # Mock only Gemini models available
             mock_get_available.return_value = {
-                "gemini-2.5-pro-preview-06-05": ProviderType.GOOGLE,
-                "gemini-2.5-flash-preview-05-20": ProviderType.GOOGLE,
+                "gemini-2.5-pro": ProviderType.GOOGLE,
+                "gemini-2.5-flash": ProviderType.GOOGLE,
             }
 
             model = ModelProviderRegistry.get_preferred_fallback_model(ToolModelCategory.EXTENDED_REASONING)
             # Should find the pro model for extended reasoning
-            assert "pro" in model or model == "gemini-2.5-pro-preview-06-05"
+            assert "pro" in model or model == "gemini-2.5-pro"
 
     def test_fast_response_with_openai(self):
         """Test FAST_RESPONSE prefers o4-mini when OpenAI is available."""
@@ -117,13 +117,13 @@ class TestModelSelection:
         with patch.object(ModelProviderRegistry, "get_available_models") as mock_get_available:
             # Mock only Gemini models available
             mock_get_available.return_value = {
-                "gemini-2.5-pro-preview-06-05": ProviderType.GOOGLE,
-                "gemini-2.5-flash-preview-05-20": ProviderType.GOOGLE,
+                "gemini-2.5-pro": ProviderType.GOOGLE,
+                "gemini-2.5-flash": ProviderType.GOOGLE,
             }
 
             model = ModelProviderRegistry.get_preferred_fallback_model(ToolModelCategory.FAST_RESPONSE)
             # Should find the flash model for fast response
-            assert "flash" in model or model == "gemini-2.5-flash-preview-05-20"
+            assert "flash" in model or model == "gemini-2.5-flash"
 
     def test_balanced_category_fallback(self):
         """Test BALANCED category uses existing logic."""
@@ -143,13 +143,13 @@ class TestModelSelection:
         with patch.object(ModelProviderRegistry, "get_available_models") as mock_get_available:
             # Mock only Gemini models available
             mock_get_available.return_value = {
-                "gemini-2.5-pro-preview-06-05": ProviderType.GOOGLE,
-                "gemini-2.5-flash-preview-05-20": ProviderType.GOOGLE,
+                "gemini-2.5-pro": ProviderType.GOOGLE,
+                "gemini-2.5-flash": ProviderType.GOOGLE,
             }
 
             model = ModelProviderRegistry.get_preferred_fallback_model()
             # Should pick a reasonable default, preferring flash for balanced use
-            assert "flash" in model or model == "gemini-2.5-flash-preview-05-20"
+            assert "flash" in model or model == "gemini-2.5-flash"
 
 
 class TestFlexibleModelSelection:
@@ -168,8 +168,8 @@ class TestFlexibleModelSelection:
             # Case 2: Mix of Gemini shorthands and full names
             {
                 "available": {
-                    "gemini-2.5-flash-preview-05-20": ProviderType.GOOGLE,
-                    "gemini-2.5-pro-preview-06-05": ProviderType.GOOGLE,
+                    "gemini-2.5-flash": ProviderType.GOOGLE,
+                    "gemini-2.5-pro": ProviderType.GOOGLE,
                 },
                 "category": ToolModelCategory.FAST_RESPONSE,
                 "expected_contains": "flash",
@@ -220,7 +220,7 @@ class TestCustomProviderFallback:
             mock_find_thinking.return_value = None
 
             model = ModelProviderRegistry.get_preferred_fallback_model(ToolModelCategory.EXTENDED_REASONING)
-            assert model == "gemini-2.5-pro-preview-06-05"
+            assert model == "gemini-2.5-pro"
 
 
 class TestAutoModeErrorMessages:
@@ -234,8 +234,8 @@ class TestAutoModeErrorMessages:
                 with patch.object(ModelProviderRegistry, "get_available_models") as mock_get_available:
                     # Mock only Gemini models available
                     mock_get_available.return_value = {
-                        "gemini-2.5-pro-preview-06-05": ProviderType.GOOGLE,
-                        "gemini-2.5-flash-preview-05-20": ProviderType.GOOGLE,
+                        "gemini-2.5-pro": ProviderType.GOOGLE,
+                        "gemini-2.5-flash": ProviderType.GOOGLE,
                     }
 
                     tool = ThinkDeepTool()
@@ -245,7 +245,7 @@ class TestAutoModeErrorMessages:
                     assert "Model parameter is required in auto mode" in result[0].text
                     # Should suggest a model suitable for extended reasoning (either full name or with 'pro')
                     response_text = result[0].text
-                    assert "gemini-2.5-pro-preview-06-05" in response_text or "pro" in response_text
+                    assert "gemini-2.5-pro" in response_text or "pro" in response_text
                     assert "(category: extended_reasoning)" in response_text
 
     @pytest.mark.asyncio
@@ -294,7 +294,7 @@ class TestFileContentPreparation:
             # Set up model context to simulate normal execution flow
             from utils.model_context import ModelContext
 
-            tool._model_context = ModelContext("gemini-2.5-pro-preview-06-05")
+            tool._model_context = ModelContext("gemini-2.5-pro")
 
             # Call the method
             content, processed_files = tool._prepare_file_content_for_prompt(["/test/file.py"], None, "test")
@@ -304,7 +304,7 @@ class TestFileContentPreparation:
             assert len(debug_calls) > 0
             debug_message = str(debug_calls[0])
             # Should mention the model being used
-            assert "gemini-2.5-pro-preview-06-05" in debug_message
+            assert "gemini-2.5-pro" in debug_message
             # Should mention file tokens (not content tokens)
             assert "file tokens" in debug_message
 
diff --git a/tests/test_prompt_regression.py b/tests/test_prompt_regression.py
index 5ad91cd..f5f47ed 100644
--- a/tests/test_prompt_regression.py
+++ b/tests/test_prompt_regression.py
@@ -31,7 +31,7 @@ class TestPromptRegression:
             return Mock(
                 content=text,
                 usage={"input_tokens": 10, "output_tokens": 20, "total_tokens": 30},
-                model_name="gemini-2.5-flash-preview-05-20",
+                model_name="gemini-2.5-flash",
                 metadata={"finish_reason": "STOP"},
             )
 
diff --git a/tests/test_providers.py b/tests/test_providers.py
index 2920c92..2b2c866 100644
--- a/tests/test_providers.py
+++ b/tests/test_providers.py
@@ -64,7 +64,7 @@ class TestModelProviderRegistry:
         """Test getting provider for a specific model"""
         ModelProviderRegistry.register_provider(ProviderType.GOOGLE, GeminiModelProvider)
 
-        provider = ModelProviderRegistry.get_provider_for_model("gemini-2.5-flash-preview-05-20")
+        provider = ModelProviderRegistry.get_provider_for_model("gemini-2.5-flash")
 
         assert provider is not None
         assert isinstance(provider, GeminiModelProvider)
@@ -95,10 +95,10 @@ class TestGeminiProvider:
         """Test getting model capabilities"""
         provider = GeminiModelProvider(api_key="test-key")
 
-        capabilities = provider.get_capabilities("gemini-2.5-flash-preview-05-20")
+        capabilities = provider.get_capabilities("gemini-2.5-flash")
 
         assert capabilities.provider == ProviderType.GOOGLE
-        assert capabilities.model_name == "gemini-2.5-flash-preview-05-20"
+        assert capabilities.model_name == "gemini-2.5-flash"
         assert capabilities.context_window == 1_048_576
         assert capabilities.supports_extended_thinking
 
@@ -106,7 +106,7 @@ class TestGeminiProvider:
         """Test getting capabilities for Pro model with thinking support"""
         provider = GeminiModelProvider(api_key="test-key")
 
-        capabilities = provider.get_capabilities("gemini-2.5-pro-preview-06-05")
+        capabilities = provider.get_capabilities("gemini-2.5-pro")
 
         assert capabilities.supports_extended_thinking
 
@@ -118,14 +118,14 @@ class TestGeminiProvider:
         assert provider.validate_model_name("pro")
 
         capabilities = provider.get_capabilities("flash")
-        assert capabilities.model_name == "gemini-2.5-flash-preview-05-20"
+        assert capabilities.model_name == "gemini-2.5-flash"
 
     def test_supports_thinking_mode(self):
         """Test thinking mode support detection"""
         provider = GeminiModelProvider(api_key="test-key")
 
-        assert provider.supports_thinking_mode("gemini-2.5-flash-preview-05-20")
-        assert provider.supports_thinking_mode("gemini-2.5-pro-preview-06-05")
+        assert provider.supports_thinking_mode("gemini-2.5-flash")
+        assert provider.supports_thinking_mode("gemini-2.5-pro")
 
     @patch("google.genai.Client")
     def test_generate_content(self, mock_client_class):
@@ -149,12 +149,12 @@ class TestGeminiProvider:
         provider = GeminiModelProvider(api_key="test-key")
 
         response = provider.generate_content(
-            prompt="Test prompt", model_name="gemini-2.5-flash-preview-05-20", temperature=0.7
+            prompt="Test prompt", model_name="gemini-2.5-flash", temperature=0.7
         )
 
         assert isinstance(response, ModelResponse)
         assert response.content == "Generated content"
-        assert response.model_name == "gemini-2.5-flash-preview-05-20"
+        assert response.model_name == "gemini-2.5-flash"
         assert response.provider == ProviderType.GOOGLE
         assert response.usage["input_tokens"] == 10
         assert response.usage["output_tokens"] == 20
diff --git a/tests/test_thinking_modes.py b/tests/test_thinking_modes.py
index 2eb3206..92969b4 100644
--- a/tests/test_thinking_modes.py
+++ b/tests/test_thinking_modes.py
@@ -418,7 +418,7 @@ class TestThinkingModes:
         from providers.gemini import GeminiModelProvider
 
         provider = GeminiModelProvider(api_key="test-key")
-        flash_model = "gemini-2.5-flash-preview-05-20"
+        flash_model = "gemini-2.5-flash"
         flash_max_tokens = 24576
 
         expected_budgets = {
diff --git a/tools/base.py b/tools/base.py
index 6440198..310c1fc 100644
--- a/tools/base.py
+++ b/tools/base.py
@@ -328,7 +328,7 @@ class BaseTool(ABC):
                 logging.debug(f"Failed to add custom models to enum: {e}")
 
         # Note: MODEL_CAPABILITIES_DESC already includes both short aliases (e.g., "flash", "o3")
-        # and full model names (e.g., "gemini-2.5-flash-preview-05-20") as keys
+        # and full model names (e.g., "gemini-2.5-flash") as keys
 
         # Remove duplicates while preserving order
         seen = set()
@@ -1097,7 +1097,7 @@ When recommending searches, be specific about what information you need and why
                 "status": "error",
                 "content": (
                     f"Image support not available: Model '{model_name}' does not support image processing. "
-                    f"Please use a vision-capable model such as 'gemini-2.5-flash-preview-05-20', 'o3', "
+                    f"Please use a vision-capable model such as 'gemini-2.5-flash', 'o3', "
                     f"or 'claude-3-opus' for image analysis tasks."
                 ),
                 "content_type": "text",
diff --git a/tools/listmodels.py b/tools/listmodels.py
index a7a932d..c1ee10c 100644
--- a/tools/listmodels.py
+++ b/tools/listmodels.py
@@ -80,8 +80,8 @@ class ListModelsTool(BaseTool):
                 "name": "Google Gemini",
                 "env_key": "GEMINI_API_KEY",
                 "models": {
-                    "flash": "gemini-2.5-flash-preview-05-20",
-                    "pro": "gemini-2.5-pro-preview-06-05",
+                    "flash": "gemini-2.5-flash",
+                    "pro": "gemini-2.5-pro",
                 },
             },
             "openai": {
diff --git a/utils/conversation_memory.py b/utils/conversation_memory.py
index 1a69480..5d4d419 100644
--- a/utils/conversation_memory.py
+++ b/utils/conversation_memory.py
@@ -159,7 +159,7 @@ class ConversationTurn(BaseModel):
         images: List of image paths referenced in this specific turn
         tool_name: Which tool generated this turn (for cross-tool tracking)
         model_provider: Provider used (e.g., "google", "openai")
-        model_name: Specific model used (e.g., "gemini-2.5-flash-preview-05-20", "o3-mini")
+        model_name: Specific model used (e.g., "gemini-2.5-flash", "o3-mini")
         model_metadata: Additional model-specific metadata (e.g., thinking mode, token usage)
     """
 
@@ -327,7 +327,7 @@ def add_turn(
         images: Optional list of images referenced in this turn
         tool_name: Name of the tool adding this turn (for attribution)
         model_provider: Provider used (e.g., "google", "openai")
-        model_name: Specific model used (e.g., "gemini-2.5-flash-preview-05-20", "o3-mini")
+        model_name: Specific model used (e.g., "gemini-2.5-flash", "o3-mini")
         model_metadata: Additional model info (e.g., thinking mode, token usage)
 
     Returns: