feat(cost): add pricing for azure_ai/gpt-image-2 and gpt-image-to-image

Adds pricing entries for new Azure AI image models.

Prices:
- Text Input: /M tokens
- Image Input: /M tokens
- Image Output: /M tokens

Closes #26765

litellm/model_prices_and_context_window_backup.json

@@ -4123,6 +4123,38 @@
         "supports_xhigh_reasoning_effort": true,
         "supports_minimal_reasoning_effort": false
     },
+    "azure_ai/gpt-image-2": {
+        "cache_read_input_image_token_cost": 2e-06,
+        "cache_read_input_token_cost": 1.25e-06,
+        "input_cost_per_token": 5e-06,
+        "input_cost_per_image_token": 8e-06,
+        "litellm_provider": "azure_ai",
+        "mode": "image_generation",
+        "output_cost_per_token": 1e-05,
+        "output_cost_per_image_token": 3e-05,
+        "supported_endpoints": [
+            "/v1/images/generations",
+            "/v1/images/edits"
+        ],
+        "supports_vision": true,
+        "supports_pdf_input": true
+    },
+    "azure_ai/gpt-image-to-image": {
+        "cache_read_input_image_token_cost": 2e-06,
+        "cache_read_input_token_cost": 1.25e-06,
+        "input_cost_per_token": 5e-06,
+        "input_cost_per_image_token": 8e-06,
+        "litellm_provider": "azure_ai",
+        "mode": "image_generation",
+        "output_cost_per_token": 1e-05,
+        "output_cost_per_image_token": 3e-05,
+        "supported_endpoints": [
+            "/v1/images/generations",
+            "/v1/images/edits"
+        ],
+        "supports_vision": true,
+        "supports_pdf_input": true
+    },
     "azure_ai/model_router": {
         "deprecation_date": "2027-05-20",
         "input_cost_per_token": 1.4e-07,

model_prices_and_context_window.json

@@ -4123,6 +4123,38 @@
         "supports_xhigh_reasoning_effort": true,
         "supports_minimal_reasoning_effort": false
     },
+    "azure_ai/gpt-image-2": {
+        "cache_read_input_image_token_cost": 2e-06,
+        "cache_read_input_token_cost": 1.25e-06,
+        "input_cost_per_token": 5e-06,
+        "input_cost_per_image_token": 8e-06,
+        "litellm_provider": "azure_ai",
+        "mode": "image_generation",
+        "output_cost_per_token": 1e-05,
+        "output_cost_per_image_token": 3e-05,
+        "supported_endpoints": [
+            "/v1/images/generations",
+            "/v1/images/edits"
+        ],
+        "supports_vision": true,
+        "supports_pdf_input": true
+    },
+    "azure_ai/gpt-image-to-image": {
+        "cache_read_input_image_token_cost": 2e-06,
+        "cache_read_input_token_cost": 1.25e-06,
+        "input_cost_per_token": 5e-06,
+        "input_cost_per_image_token": 8e-06,
+        "litellm_provider": "azure_ai",
+        "mode": "image_generation",
+        "output_cost_per_token": 1e-05,
+        "output_cost_per_image_token": 3e-05,
+        "supported_endpoints": [
+            "/v1/images/generations",
+            "/v1/images/edits"
+        ],
+        "supports_vision": true,
+        "supports_pdf_input": true
+    },
     "azure_ai/model_router": {
         "deprecation_date": "2027-05-20",
         "input_cost_per_token": 1.4e-07,

tests/test_litellm/test_utils.py

@@ -4116,6 +4116,37 @@ class TestValidateAndFixThinkingParam:
         assert validate_and_fix_thinking_param(thinking=False) is None
 
 
+def test_azure_ai_gpt_image_models_in_cost_map():
+    """
+    Test that azure_ai/gpt-image-2 and azure_ai/gpt-image-to-image entries 
+    are correctly configured in model_prices_and_context_window.json.
+
+    Prices:
+    - Text Input:  $5/M tokens
+    - Image Input: $8/M tokens
+    - Image Output: $30/M tokens
+
+    Closes https://github.com/BerriAI/litellm/issues/26765
+    """
+    import json
+    from pathlib import Path
+
+    json_path = Path(__file__).parents[2] / "model_prices_and_context_window.json"
+    with open(json_path) as f:
+        model_cost = json.load(f)
+
+    for key in ["azure_ai/gpt-image-2", "azure_ai/gpt-image-to-image"]:
+        info = model_cost.get(key)
+        assert info is not None, f"{key} missing from model_prices_and_context_window.json"
+        assert info["litellm_provider"] == "azure_ai"
+        assert info["mode"] == "image_generation"
+        assert info["input_cost_per_token"] == 5e-06
+        assert info["input_cost_per_image_token"] == 8e-06
+        assert info["output_cost_per_image_token"] == 3e-05
+        assert info["cache_read_input_image_token_cost"] == 2e-06
+        assert info["supports_vision"] is True
+
+
 def test_deepseek_v4_models_in_cost_map():
     """
     Test that deepseek-v4-flash and deepseek-v4-pro entries are correctly