fix(perplexity): bill search queries at the per-request price, not 1/1000 of it

litellm/llms/perplexity/cost_calculator.py

@@ -98,10 +98,11 @@ def _safe_float_cast(
     if num_search_queries > 0 and search_cost_value is not None:
         # Handle both dict and float formats
         if isinstance(search_cost_value, dict):
-            # Use the "low" size as default - tests expect 0.005 / 1000
-            search_cost_per_query = (
-                _safe_float_cast(search_cost_value.get("search_context_size_low", 0))
-                / 1000
+            # search_context_cost_per_query stores the per-request price in USD
+            # (e.g. sonar low = $0.005/request). Use it directly, matching the
+            # gemini cost calculator which reads the same field per request.
+            search_cost_per_query = _safe_float_cast(
+                search_cost_value.get("search_context_size_low", 0)
             )
         else:
             search_cost_per_query = _safe_float_cast(search_cost_value)

tests/test_litellm/llms/perplexity/test_perplexity_cost_calculator.py

@@ -9,7 +9,7 @@
 import math
 import os
 import sys
-from unittest.mock import Mock, patch
+from unittest.mock import patch
 
 import pytest
 
@@ -120,10 +120,10 @@ def test_search_queries_cost_calculation(self):
         # Expected costs:
         # Input: 100 tokens * $2e-6 = $0.0002
         # Output: 50 tokens * $8e-6 = $0.0004
-        # Search: 3 queries * ($0.005 / 1000) = $0.000015
-        # Total completion cost: $0.000415
+        # Search: 3 queries * $0.005 per request = $0.015
+        # Total completion cost: $0.0154
         expected_prompt_cost = 100 * 2e-6
-        expected_completion_cost = (50 * 8e-6) + (3 / 1000 * 0.005)
+        expected_completion_cost = (50 * 8e-6) + (3 * 0.005)
 
         assert math.isclose(prompt_cost, expected_prompt_cost, rel_tol=1e-6)
         assert math.isclose(completion_cost, expected_completion_cost, rel_tol=1e-6)
@@ -195,10 +195,10 @@ def test_comprehensive_cost_calculation(self):
         # Total prompt cost                        = $0.00026
         # Output (text): (50 - 15) tokens  * $8e-6 = $0.00028
         # Reasoning:      15 tokens        * $3e-6 = $0.000045
-        # Search:          2 queries * ($0.005 / 1000) = $0.00001
-        # Total completion cost                    = $0.000335
+        # Search:          2 queries * $0.005 per request = $0.01
+        # Total completion cost                    = $0.010325
         expected_prompt_cost = (100 * 2e-6) + (30 * 2e-6)
-        expected_completion_cost = ((50 - 15) * 8e-6) + (15 * 3e-6) + (2 / 1000 * 0.005)
+        expected_completion_cost = ((50 - 15) * 8e-6) + (15 * 3e-6) + (2 * 0.005)
 
         assert math.isclose(prompt_cost, expected_prompt_cost, rel_tol=1e-6)
         assert math.isclose(completion_cost, expected_completion_cost, rel_tol=1e-6)
@@ -311,7 +311,7 @@ def test_integration_with_completion_cost_function(self):
         # Calculate expected total cost (reasoning is a subset of completion_tokens)
         expected_prompt_cost = (100 * 2e-6) + (15 * 2e-6)  # Input + citation
         expected_completion_cost = (
-            ((50 - 10) * 8e-6) + (10 * 3e-6) + (1 / 1000 * 0.005)
+            ((50 - 10) * 8e-6) + (10 * 3e-6) + (1 * 0.005)
         )  # Output (text) + reasoning + search
         expected_total = expected_prompt_cost + expected_completion_cost
 
@@ -361,7 +361,7 @@ def test_cost_calculation_combinations(
         expected_completion_cost = (
             ((50 - reasoning_tokens) * 8e-6)
             + (reasoning_tokens * 3e-6)
-            + (search_queries / 1000 * 0.005)
+            + (search_queries * 0.005)
         )
 
         assert math.isclose(prompt_cost, expected_prompt_cost, rel_tol=1e-6)

tests/test_litellm/llms/perplexity/test_perplexity_integration.py

@@ -9,7 +9,6 @@
 import math
 import os
 import sys
-from unittest.mock import Mock, patch
 
 import pytest
 
@@ -106,8 +105,8 @@ def test_end_to_end_cost_calculation_with_transformation(self):
 
         expected_prompt_cost = (100 * 2e-6) + (citation_tokens * 2e-6)
         expected_completion_cost = (
-            ((50 - 10) * 8e-6) + (10 * 3e-6) + (2 / 1000 * 0.005)
-        )
+            ((50 - 10) * 8e-6) + (10 * 3e-6) + (2 * 0.005)
+        )  # Output (text) + reasoning + search
         expected_total = expected_prompt_cost + expected_completion_cost
 
         assert math.isclose(total_cost, expected_total, rel_tol=1e-6)
@@ -152,8 +151,8 @@ def test_main_cost_calculator_integration(self):
 
         expected_prompt_cost = (200 * 2e-6) + (40 * 2e-6)
         expected_completion_cost = (
-            ((100 - 25) * 8e-6) + (25 * 3e-6) + (3 / 1000 * 0.005)
-        )
+            ((100 - 25) * 8e-6) + (25 * 3e-6) + (3 * 0.005)
+        )  # Output (text) + reasoning + search
 
         assert math.isclose(prompt_cost, expected_prompt_cost, rel_tol=1e-6)
         assert math.isclose(completion_cost_val, expected_completion_cost, rel_tol=1e-6)
@@ -262,9 +261,9 @@ def test_high_volume_cost_calculation(self):
 
         expected_prompt_cost = (50000 * 2e-6) + (5000 * 2e-6)
         expected_completion_cost = (
-            ((25000 - 10000) * 8e-6) + (10000 * 3e-6) + (100 / 1000 * 0.005)
-        )
-        expected_total = expected_prompt_cost + expected_completion_cost
+            ((25000 - 10000) * 8e-6) + (10000 * 3e-6) + (100 * 0.005)
+        )  # $0.65
+        expected_total = expected_prompt_cost + expected_completion_cost  # $0.76
 
         assert math.isclose(total_cost, expected_total, rel_tol=1e-6)
         assert total_cost > 0.25
@@ -326,7 +325,7 @@ def test_case_insensitive_provider_matching(self, provider_name):
 
         # Should calculate costs correctly
         expected_prompt_cost = (100 * 2e-6) + (10 * 2e-6)
-        expected_completion_cost = (50 * 8e-6) + (1 / 1000 * 0.005)
+        expected_completion_cost = (50 * 8e-6) + (1 * 0.005)
 
         assert math.isclose(prompt_cost, expected_prompt_cost, rel_tol=1e-6)
         assert math.isclose(completion_cost_val, expected_completion_cost, rel_tol=1e-6)