fix(model_info): guard backfill by mode, drop provider key, tighten claude major regex

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

litellm/litellm_core_utils/fallback_generalizations.py

@@ -191,7 +191,10 @@ class _FallbackGeneralizations:
         matched = tuple(rule.model_info for rule in self.backfill_rules if rule.pattern.search(model) is not None)
         if not matched:
             return None
-        return {key: value for model_info in matched for key, value in model_info.items()}
+        backfill: Final = {
+            key: value for model_info in matched for key, value in model_info.items() if key != PROVIDER_KEY
+        }
+        return backfill or None
 
 
 _registry: Final = _FallbackGeneralizations()

litellm/model_prices_and_context_window_backup.json

@@ -57716,9 +57716,9 @@
             },
             {
                 "name": "claude-adaptive-thinking",
-                "pattern": "claude-[a-z]+-(?:4[-._](?:[6-9]|[1-9]\\d)(?!\\d)|(?:[5-9]|[1-9]\\d)(?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
+                "pattern": "claude-[a-z]+-(?:4[-._](?:[6-9]|[1-9]\\d)(?!\\d)|[5-9](?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
                 "backfill_exact_entries": true,
-                "description": "Claude at version 4.6 or higher, in any id shape that contains claude-<family>-: minors 4.6 through 4.99, any later major-minor, and bare 5+ majors so a new family shaped like claude-fable-5 matches. Requiring the claude- prefix keeps non-Claude names such as team-sonnet-5-1 out. The minor is capped at two digits so an 8-digit date suffix such as claude-opus-4-20250514 is never read as a >= 4.6 minor. Turns on adaptive thinking for new versions and new families with no code change.",
+                "description": "Claude at version 4.6 or higher, in any id shape that contains claude-<family>-: minors 4.6 through 4.99, any later major-minor, and bare 5+ majors so a new family shaped like claude-fable-5 matches. Requiring the claude- prefix keeps non-Claude names such as team-sonnet-5-1 out. The minor is capped at two digits so an 8-digit date suffix such as claude-opus-4-20250514 is never read as a >= 4.6 minor. Two-digit majors are deliberately not matched so ids like claude-opus-41 (4.1) are not read as major 41. Turns on adaptive thinking for new versions and new families with no code change.",
                 "model_info": {
                     "supports_adaptive_thinking": true
                 }
@@ -57742,9 +57742,9 @@
             },
             {
                 "name": "claude-mid-conversation-system",
-                "pattern": "claude-[a-z]+-(?:4[-._](?:[89]|[1-9]\\d)(?!\\d)|(?:[5-9]|[1-9]\\d)(?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
+                "pattern": "claude-[a-z]+-(?:4[-._](?:[89]|[1-9]\\d)(?!\\d)|[5-9](?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
                 "backfill_exact_entries": true,
-                "description": "Claude at version 4.8 or higher, in any id shape that contains claude-<family>-: minors 4.8 through 4.99, any later major-minor, and bare 5+ majors so a new family like claude-fable-5 matches. Anthropic introduced mid-conversation system messages with Opus 4.8 and every newer Claude keeps them; 4.7 and below reject the system role inside messages.",
+                "description": "Claude at version 4.8 or higher, in any id shape that contains claude-<family>-: minors 4.8 through 4.99, any later major-minor, and bare 5+ majors so a new family like claude-fable-5 matches. Two-digit majors are deliberately not matched so ids like claude-opus-41 (4.1) are not read as major 41. Anthropic introduced mid-conversation system messages with Opus 4.8 and every newer Claude keeps them; 4.7 and below reject the system role inside messages.",
                 "model_info": {
                     "supports_mid_conversation_system": true
                 }

litellm/utils.py

@@ -255,6 +255,7 @@ from litellm.types.utils import (
 )
 
 _CALL_TYPE_ENUM_MAP: Final[dict] = {ct.value: ct for ct in CallTypes}
+_BACKFILL_MODES: Final = frozenset({"chat", "responses"})
 
 # +-----------------------------------------------+
 # |                                               |
@@ -5820,7 +5821,7 @@ def _get_model_info_helper(
                     ):
                         _model_info = None
 
-            if _model_info is not None and key is not None:
+            if _model_info is not None and key is not None and _model_info.get("mode", "chat") in _BACKFILL_MODES:
                 backfill: Final = match_backfill_generalizations(key)
                 if backfill is not None:
                     _model_info = {**{k: v for k, v in backfill.items() if k not in _model_info}, **_model_info}

model_prices_and_context_window.json

@@ -57716,9 +57716,9 @@
             },
             {
                 "name": "claude-adaptive-thinking",
-                "pattern": "claude-[a-z]+-(?:4[-._](?:[6-9]|[1-9]\\d)(?!\\d)|(?:[5-9]|[1-9]\\d)(?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
+                "pattern": "claude-[a-z]+-(?:4[-._](?:[6-9]|[1-9]\\d)(?!\\d)|[5-9](?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
                 "backfill_exact_entries": true,
-                "description": "Claude at version 4.6 or higher, in any id shape that contains claude-<family>-: minors 4.6 through 4.99, any later major-minor, and bare 5+ majors so a new family shaped like claude-fable-5 matches. Requiring the claude- prefix keeps non-Claude names such as team-sonnet-5-1 out. The minor is capped at two digits so an 8-digit date suffix such as claude-opus-4-20250514 is never read as a >= 4.6 minor. Turns on adaptive thinking for new versions and new families with no code change.",
+                "description": "Claude at version 4.6 or higher, in any id shape that contains claude-<family>-: minors 4.6 through 4.99, any later major-minor, and bare 5+ majors so a new family shaped like claude-fable-5 matches. Requiring the claude- prefix keeps non-Claude names such as team-sonnet-5-1 out. The minor is capped at two digits so an 8-digit date suffix such as claude-opus-4-20250514 is never read as a >= 4.6 minor. Two-digit majors are deliberately not matched so ids like claude-opus-41 (4.1) are not read as major 41. Turns on adaptive thinking for new versions and new families with no code change.",
                 "model_info": {
                     "supports_adaptive_thinking": true
                 }
@@ -57742,9 +57742,9 @@
             },
             {
                 "name": "claude-mid-conversation-system",
-                "pattern": "claude-[a-z]+-(?:4[-._](?:[89]|[1-9]\\d)(?!\\d)|(?:[5-9]|[1-9]\\d)(?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
+                "pattern": "claude-[a-z]+-(?:4[-._](?:[89]|[1-9]\\d)(?!\\d)|[5-9](?!\\d)(?:[-._]\\d{1,2}(?!\\d))?)",
                 "backfill_exact_entries": true,
-                "description": "Claude at version 4.8 or higher, in any id shape that contains claude-<family>-: minors 4.8 through 4.99, any later major-minor, and bare 5+ majors so a new family like claude-fable-5 matches. Anthropic introduced mid-conversation system messages with Opus 4.8 and every newer Claude keeps them; 4.7 and below reject the system role inside messages.",
+                "description": "Claude at version 4.8 or higher, in any id shape that contains claude-<family>-: minors 4.8 through 4.99, any later major-minor, and bare 5+ majors so a new family like claude-fable-5 matches. Two-digit majors are deliberately not matched so ids like claude-opus-41 (4.1) are not read as major 41. Anthropic introduced mid-conversation system messages with Opus 4.8 and every newer Claude keeps them; 4.7 and below reject the system role inside messages.",
                 "model_info": {
                     "supports_mid_conversation_system": true
                 }

tests/test_litellm/litellm_core_utils/test_fallback_generalizations.py

@@ -139,6 +139,18 @@ def test_backfill_requires_per_rule_opt_in(restore_generalizations):
     )
     assert match_backfill_generalizations("acme-1") is None
 
+    restore_generalizations(
+        [
+            {
+                "name": "mixed",
+                "pattern": r"^acme-",
+                "backfill_exact_entries": True,
+                "model_info": {"litellm_provider": "openai", "supports_vision": True},
+            }
+        ]
+    )
+    assert match_backfill_generalizations("acme-1") == {"supports_vision": True}
+
     restore_generalizations(
         [
             {
@@ -355,6 +367,12 @@ def test_exact_entries_backfill_only_missing_fields(restore_generalizations, mon
                 "litellm_provider": "openai",
                 "mode": "chat",
             },
+            "acme-image": {
+                "input_cost_per_token": 5e-6,
+                "output_cost_per_token": 6e-6,
+                "litellm_provider": "openai",
+                "mode": "image_generation",
+            },
         },
     )
     restore_generalizations(
@@ -379,6 +397,9 @@ def test_exact_entries_backfill_only_missing_fields(restore_generalizations, mon
     assert bare["input_cost_per_token"] == 3e-6
     assert bare["key"] == "acme-bare"
 
+    image = litellm.get_model_info("acme-image", custom_llm_provider="openai")
+    assert image.get("supports_reasoning") is None
+
     restore_generalizations(
         [{"name": "acme-backfill", "pattern": r"^acme-", "model_info": {"supports_reasoning": True, "max_tokens": 5}}]
     )
@@ -503,6 +524,18 @@ def test_shipped_version_boundaries(shipped_cost_map, model, provider, adaptive,
     assert info.get("supports_mid_conversation_system") is mid_conversation, model
 
 
+def test_shipped_claude_version_regex_excludes_undelimited_41(shipped_cost_map):
+    unmatched = match_capability_generalizations("github_copilot/claude-opus-41")
+    assert unmatched is None or "supports_adaptive_thinking" not in unmatched
+    assert unmatched is None or "supports_mid_conversation_system" not in unmatched
+
+    for model in ("claude-opus-5", "claude-sonnet-4-8"):
+        matched = match_capability_generalizations(model)
+        assert matched is not None
+        assert matched["supports_adaptive_thinking"] is True
+        assert matched["supports_mid_conversation_system"] is True
+
+
 def test_shipped_rules_cover_new_families_like_fable_at_5_plus(shipped_cost_map):
     """Both version gates accept any claude-<family>- id at major 5 or higher, bare
     major or major-minor, so a new family shaped like claude-fable-5 gets adaptive
@@ -833,6 +866,17 @@ def test_shipped_openai_reasoning_rule_backfills_mapped_entries(shipped_cost_map
     assert info["input_cost_per_token"] == raw_entry.get("input_cost_per_token", 0)
 
 
+def test_shipped_openai_reasoning_rule_skips_non_text_modes(shipped_cost_map):
+    model = "gemini/deep-research-pro-preview-12-2025"
+    assert model in litellm.model_cost
+    raw_entry = litellm.model_cost[model]
+    assert "supports_reasoning" not in raw_entry
+    assert raw_entry["mode"] == "image_generation"
+
+    info = litellm.get_model_info("deep-research-pro-preview-12-2025", custom_llm_provider="gemini")
+    assert info.get("supports_reasoning") is None
+
+
 def test_shipped_claude_thinking_rules_backfill_without_family_limits(shipped_cost_map):
     model = "perplexity/anthropic/claude-sonnet-4-6"
     assert model in litellm.model_cost