diff --git a/agent/model_metadata.py b/agent/model_metadata.py
index d3223002eed0..03c05177b1f6 100644
--- a/agent/model_metadata.py
+++ b/agent/model_metadata.py
@@ -320,14 +320,12 @@ DEFAULT_CONTEXT_LENGTHS = {
     "kimi": 262144,
     # Upstage Solar — api.upstage.ai/v1/models does not return context_length,
     # so these fallbacks keep token budgeting / compression from probing down
-    # to the 128k default. Substring matching is longest-first, so the versioned
-    # ids win over the "solar-pro" rolling-alias entry, which in turn covers
-    # future solar-pro* releases at the Pro context size.
+    # to the 128k default. Ids are matched longest-first, so dated variants
+    # (e.g. solar-pro3-250127) resolve via their family prefix.
     # Sources: Solar Pro 3 = 128K, Solar Pro 2 = 64K, Solar Mini = 32K,
     # Solar Open 2 = 256K.
     "solar-open2": 262144,  # 256K
     "solar-pro3": 131072,
-    "solar-pro": 131072,  # rolling alias → latest Solar Pro (currently pro3)
     "solar-pro2": 65536,
     "solar-mini": 32768,
     # Tencent — Hy3 Preview (Hunyuan) with 256K context window.
diff --git a/plugins/model-providers/upstage/__init__.py b/plugins/model-providers/upstage/__init__.py
index 982e741f9e86..7bc202d4bf6a 100644
--- a/plugins/model-providers/upstage/__init__.py
+++ b/plugins/model-providers/upstage/__init__.py
@@ -33,8 +33,8 @@ def _model_supports_reasoning(model: str | None) -> bool:
     other (incl. future) Solar model as reasoning-capable.
 
     ``None``/empty model → True: the provider default (``fallback_models[0]``,
-    the ``solar-pro`` rolling alias) is reasoning-capable, so an unset model
-    gets the same default-on behaviour.
+    ``solar-pro3``) is reasoning-capable, so an unset model gets the same
+    default-on behaviour.
     """
     m = (model or "").strip().lower()
     return not any(marker in m for marker in _NON_REASONING_MODEL_MARKERS)
@@ -103,10 +103,8 @@ upstage = UpstageProfile(
     base_url="https://api.upstage.ai/v1",
     auth_type="api_key",
     # default_aux_model left empty → auxiliary side tasks use the main model.
-    # entry [0] is the setup default. solar-pro is a rolling alias for the
-    # latest Solar Pro, so the default tracks the current flagship.
+    # entry [0] is the setup default — solar-pro3, the current Solar Pro flagship.
     fallback_models=(
-        "solar-pro",
         "solar-pro3",
     ),
 )
diff --git a/tests/plugins/model_providers/test_upstage_profile.py b/tests/plugins/model_providers/test_upstage_profile.py
index e78c2326ee11..f37971f2f791 100644
--- a/tests/plugins/model_providers/test_upstage_profile.py
+++ b/tests/plugins/model_providers/test_upstage_profile.py
@@ -50,13 +50,12 @@ class TestUpstageProfile:
         # catalog — Mini is capable but not agentic, so it's never promoted as a
         # default. Live /v1/models still surfaces everything when a key is set.
         assert upstage_profile.fallback_models == (
-            "solar-pro",
             "solar-pro3",
         )
 
-    def test_default_model_is_solar_pro(self, upstage_profile):
+    def test_default_model_is_solar_pro3(self, upstage_profile):
         # Entry [0] is the setup default (get_default_model_for_provider).
-        assert upstage_profile.fallback_models[0] == "solar-pro"
+        assert upstage_profile.fallback_models[0] == "solar-pro3"
 
     def test_aux_model_left_empty(self, upstage_profile):
         # Unset → auxiliary side tasks fall back to the user's main model.
@@ -163,7 +162,7 @@ class TestUpstageReasoning:
 
     def test_none_model_defaults_to_reasoning(self, upstage_profile):
         # No model in context → treated as reasoning-capable, consistent with
-        # the provider default (fallback_models[0] == "solar-pro").
+        # the provider default (fallback_models[0] == "solar-pro3").
         _, top_level = upstage_profile.build_api_kwargs_extras(model=None)
         assert top_level == {"reasoning_effort": "medium"}