diff --git a/config.example.yaml b/config.example.yaml
index 296a859..bfd84ab 100644
--- a/config.example.yaml
+++ b/config.example.yaml
@@ -30,7 +30,7 @@ llm:
     - "openai"
   gemini:
     api_key: ""  # or set LOGO_LLM_GEMINI_API_KEY env var
-    model: "gemini-2.5-flash"  # free tier — 15 req/min
+    model: "gemini-2.5-flash-lite"  # free tier — 1000 req/day vs flash's 20/day
   anthropic:
     api_key: ""  # or set LOGO_LLM_ANTHROPIC_API_KEY env var
     model: "claude-sonnet-4-5-20250929"
diff --git a/internal/config/config.go b/internal/config/config.go
index 668c183..01972dd 100644
--- a/internal/config/config.go
+++ b/internal/config/config.go
@@ -96,7 +96,9 @@ func Load(configPath string) (*Config, error) {
 	v.SetDefault("llm.provider_order", []string{"gemini", "anthropic", "openai"})
 	v.SetDefault("llm.anthropic.model", "claude-sonnet-4-5-20250929")
 	v.SetDefault("llm.openai.model", "gpt-4o")
-	v.SetDefault("llm.gemini.model", "gemini-2.5-flash")
+	// flash-lite gets 1000 free req/day vs flash's 20 — and Wikidata covers
+	// the major-company case now, so LLM is mostly a long-tail fallback.
+	v.SetDefault("llm.gemini.model", "gemini-2.5-flash-lite")
 	v.SetDefault("llm.rate_per_minute", 10)
 	v.SetDefault("github.repos", []string{
 		"davidepalazzo/ticker-logos",
diff --git a/internal/service/logo_service.go b/internal/service/logo_service.go
index 5821388..89d679e 100644
--- a/internal/service/logo_service.go
+++ b/internal/service/logo_service.go
@@ -145,8 +145,12 @@ func (s *LogoService) acquire(ctx context.Context, symbol, companyName string) (
 			)
 			return result, nil
 		}
-		s.logger.Debug("wikidata provider miss",
+		// Info (not Debug) so we can see in production whether Wikidata was
+		// even reachable for a given ticker and why it bailed (missing
+		// company_name vs no Q-entity vs no P154 vs download fail).
+		s.logger.Info("wikidata provider miss",
 			zap.String("symbol", symbol),
+			zap.String("company_name", companyName),
 			zap.Error(err),
 		)
 	}