feat(google-vertex): update model YAMLs [bot]

providers/google-vertex/anthropic/claude-haiku-4-5@20251001.yaml

@@ -7,6 +7,14 @@ costs:
       output_cost_per_token: 0.0000055
       output_cost_per_token_batches: 0.00000275
       region: us-east5
+    - cache_creation_input_token_cost: 0.000001375
+      cache_creation_input_token_cost_per_hour: 0.0000022
+      cache_read_input_token_cost: 1.1e-7
+      input_cost_per_token: 0.0000011
+      input_cost_per_token_batches: 5.5e-7
+      output_cost_per_token: 0.0000055
+      output_cost_per_token_batches: 0.00000275
+      region: us
     - cache_creation_input_token_cost: 0.00000125
       cache_creation_input_token_cost_per_hour: 0.000002
       cache_read_input_token_cost: 1e-7
@@ -23,6 +31,14 @@ costs:
       output_cost_per_token: 0.0000055
       output_cost_per_token_batches: 0.00000275
       region: europe-west1
+    - cache_creation_input_token_cost: 0.000001375
+      cache_creation_input_token_cost_per_hour: 0.0000022
+      cache_read_input_token_cost: 1.1e-7
+      input_cost_per_token: 0.0000011
+      input_cost_per_token_batches: 5.5e-7
+      output_cost_per_token: 0.0000055
+      output_cost_per_token_batches: 0.00000275
+      region: eu
     - cache_creation_input_token_cost: 0.000001375
       cache_creation_input_token_cost_per_hour: 0.0000022
       cache_read_input_token_cost: 1.1e-7
@@ -60,6 +76,8 @@ removeParams:
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5
     - https://platform.claude.com/docs/en/about-claude/pricing
+    - https://platform.claude.com/docs/en/about-claude/models/overview
+    - https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai
 status: active
 supportedModes:
     - chat

providers/google-vertex/anthropic/claude-opus-4-6.yaml

@@ -3,13 +3,17 @@ costs:
       cache_creation_input_token_cost_per_hour: 0.000011
       cache_read_input_token_cost: 5.5e-7
       input_cost_per_token: 0.0000055
+      input_cost_per_token_batches: 0.00000275
       output_cost_per_token: 0.0000275
+      output_cost_per_token_batches: 0.00001375
       region: us-east5
     - cache_creation_input_token_cost: 0.000006875
       cache_creation_input_token_cost_per_hour: 0.000011
       cache_read_input_token_cost: 5.5e-7
       input_cost_per_token: 0.0000055
+      input_cost_per_token_batches: 0.00000275
       output_cost_per_token: 0.0000275
+      output_cost_per_token_batches: 0.00001375
       region: us
     - cache_creation_input_token_cost: 0.00000625
       cache_creation_input_token_cost_per_hour: 0.00001
@@ -23,19 +27,25 @@ costs:
       cache_creation_input_token_cost_per_hour: 0.000011
       cache_read_input_token_cost: 5.5e-7
       input_cost_per_token: 0.0000055
+      input_cost_per_token_batches: 0.00000275
       output_cost_per_token: 0.0000275
+      output_cost_per_token_batches: 0.00001375
       region: europe-west1
     - cache_creation_input_token_cost: 0.000006875
       cache_creation_input_token_cost_per_hour: 0.000011
       cache_read_input_token_cost: 5.5e-7
       input_cost_per_token: 0.0000055
+      input_cost_per_token_batches: 0.00000275
       output_cost_per_token: 0.0000275
+      output_cost_per_token_batches: 0.00001375
       region: eu
     - cache_creation_input_token_cost: 0.000006875
       cache_creation_input_token_cost_per_hour: 0.000011
       cache_read_input_token_cost: 5.5e-7
       input_cost_per_token: 0.0000055
+      input_cost_per_token_batches: 0.00000275
       output_cost_per_token: 0.0000275
+      output_cost_per_token_batches: 0.00001375
       region: asia-southeast1
 features:
     - function_calling

providers/google-vertex/anthropic/claude-sonnet-4-5.yaml

@@ -156,6 +156,7 @@ sources:
     - https://platform.claude.com/docs/en/about-claude/pricing
     - https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-claude
+    - https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/partner-models/claude/sonnet-4-5
     - https://cloud.google.com/blog/products/ai-machine-learning/multi-region-endpoints-for-claude-available-on-vertex-ai
 status: active
 supportedModes:

providers/google-vertex/dandelin/vilt-b32-finetuned-vqa.yaml

@@ -11,5 +11,6 @@ sources:
     - https://huggingface.co/dandelin/vilt-b32-finetuned-vqa
     - https://console.cloud.google.com/vertex-ai/publishers/dandelin/model-garden/vilt-b32-finetuned-vqa
     - https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vilt_vqa.ipynb
+status: active
 supportedModes:
     - unknown

providers/google-vertex/deepseek-ai/deepseek-ocr-maas.yaml

@@ -1,8 +1,10 @@
 costs:
     - input_cost_per_page: 0.0003
       input_cost_per_token: 3e-7
-      output_cost_per_token: 0.0000012
-      region: global # docs specify us-central1 but the model is only available via global inference
+      output_cost_per_token: 1.2e-6
+      region: global
+deprecationDate: "2026-07-21"
+isDeprecated: true
 limits:
     context_window: 8192
     max_output_tokens: 8192
@@ -22,9 +24,10 @@ params:
       maxValue: 8192
       minValue: 1
 provisioning: serverless
+retirementDate: "2026-10-21"
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek/deepseek-ocr
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek
-status: active
+status: deprecated
 supportedModes:
     - ocr

providers/google-vertex/deepseek-ai/deepseek-r1-0528-maas.yaml

@@ -1,14 +1,16 @@
 costs:
-    - input_cost_per_token: 0.00000135
+    - input_cost_per_token: 1.35e-6
       input_cost_per_token_batches: 6.75e-7
-      output_cost_per_token: 0.0000054
-      output_cost_per_token_batches: 0.0000027
+      output_cost_per_token: 5.4e-6
+      output_cost_per_token_batches: 2.7e-6
       region: us-central1
+deprecationDate: "2026-07-21"
 features:
     - function_calling
     - structured_output
     - system_messages
     - tool_choice
+isDeprecated: true
 limits:
     context_window: 163840
     max_output_tokens: 32768
@@ -22,10 +24,11 @@ modalities:
 mode: chat
 model: deepseek-ai/deepseek-r1-0528-maas
 provisioning: serverless
+retirementDate: "2026-10-21"
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek/r1-0528
-status: active
+status: deprecated
 supportedModes:
     - chat
 thinking: true

providers/google-vertex/deepseek-ai/deepseek-v3-1.yaml

@@ -2,13 +2,15 @@ costs:
     - cache_read_input_token_cost: 6e-8
       input_cost_per_token: 6e-7
       input_cost_per_token_batches: 3e-7
-      output_cost_per_token: 0.0000017
+      output_cost_per_token: 1.7e-6
       output_cost_per_token_batches: 8.5e-7
       region: us-central1
+deprecationDate: "2026-07-21"
 features:
     - function_calling
     - structured_output
     - prompt_caching
+isDeprecated: true
 limits:
     context_window: 163840
     max_output_tokens: 32768
@@ -21,10 +23,11 @@ modalities:
 mode: chat
 model: deepseek-ai/deepseek-v3-1
 provisioning: provisioned
+retirementDate: "2026-10-21"
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek/deepseek-v31
     - https://cloud.google.com/vertex-ai/generative-ai/pricing#deepseek-models
-status: active
+status: deprecated
 supportedModes:
     - chat
 thinking: true

providers/google-vertex/deepseek-ai/deepseek-v3-2.yaml

@@ -2,13 +2,15 @@ costs:
     - cache_read_input_token_cost: 5.6e-8
       input_cost_per_token: 5.6e-7
       input_cost_per_token_batches: 2.8e-7
-      output_cost_per_token: 0.00000168
+      output_cost_per_token: 1.68e-6
       output_cost_per_token_batches: 8.4e-7
       region: global
+deprecationDate: "2026-07-21"
 features:
     - function_calling
     - structured_output
     - prompt_caching
+isDeprecated: true
 limits:
     context_window: 163840
     max_output_tokens: 65536
@@ -26,9 +28,10 @@ params:
       maxValue: 65536
       minValue: 1
 provisioning: provisioned
+retirementDate: "2026-10-21"
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek/deepseek-v32
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek
-status: active
+status: deprecated
 supportedModes:
     - chat

providers/google-vertex/deepseek-ai/deepseek-v3.1-maas.yaml

@@ -2,15 +2,17 @@ costs:
     - cache_read_input_token_cost: 6e-8
       input_cost_per_token: 6e-7
       input_cost_per_token_batches: 3e-7
-      output_cost_per_token: 0.0000017
+      output_cost_per_token: 1.7e-6
       output_cost_per_token_batches: 8.5e-7
-      region: us-west2 # docs specify us-central1 but the model is available via us-west2 inference
+      region: us-west2
+deprecationDate: "2026-07-21"
 features:
     - function_calling
     - tool_choice
     - structured_output
     - system_messages
     - prompt_caching
+isDeprecated: true
 limits:
     context_window: 163840
     max_output_tokens: 32768
@@ -24,11 +26,12 @@ modalities:
 mode: chat
 model: deepseek-ai/deepseek-v3.1-maas
 provisioning: serverless
+retirementDate: "2026-10-21"
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek/deepseek-v31
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/deepseek
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/maas/use-open-models
-status: active
+status: deprecated
 supportedModes:
     - chat
 thinking: true

providers/google-vertex/gemini-3-pro-image-preview.yaml

@@ -1,12 +1,13 @@
 costs:
     - input_cost_per_image: 0.00112
       input_cost_per_token: 0.000002
+      input_cost_per_token_batches: 0.000001
       output_cost_per_image_token: 0.00012
       output_cost_per_token: 0.000012
+      output_cost_per_token_batches: 0.000006
       region: global
 features:
     - system_messages
-    - structured_output
 limits:
     context_window: 65536
     max_input_tokens: 65536
@@ -22,6 +23,15 @@ modalities:
         - image
 mode: image
 model: gemini-3-pro-image-preview
+params:
+    - defaultValue: 1.0
+      key: temperature
+      maxValue: 2
+      minValue: 0
+    - defaultValue: 0.95
+      key: top_p
+      maxValue: 1
+      minValue: 0
 provisioning: serverless
 removeParams:
     - tool_choice

providers/google-vertex/gemma3.yaml

@@ -6,6 +6,7 @@ limits:
     context_window: 128000
     max_input_tokens: 128000
     max_output_tokens: 128000
+    max_tokens: 128000
 modalities:
     input:
         - text

providers/google-vertex/google/gemini-2.5-flash-lite.yaml

@@ -129,8 +129,8 @@ features:
 limits:
     context_window: 1048576
     max_input_tokens: 1048576
-    max_output_tokens: 65535
-    max_tokens: 65535
+    max_output_tokens: 65536
+    max_tokens: 65536
 modalities:
     input:
         - text
@@ -149,10 +149,11 @@ params:
       key: top_p
     - defaultValue: 256
       key: max_tokens
-      maxValue: 65535
+      maxValue: 65536
 provisioning: serverless
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-flash-lite
+    - https://ai.google.dev/gemini-api/docs/models/gemini-2.5-flash-lite
 status: active
 supportedModes:
     - chat

providers/google-vertex/google/gemini-2.5-pro-tts.yaml

@@ -91,6 +91,6 @@ sources:
     - https://ai.google.dev/gemini-api/docs/models
     - https://ai.google.dev/gemini-api/docs/pricing
     - https://docs.cloud.google.com/text-to-speech/docs/gemini-tts
-status: active
+status: preview
 supportedModes:
     - text_to_speech

providers/google-vertex/google/gemini-3.1-pro-preview.yaml

@@ -1,5 +1,6 @@
 costs:
-    - cache_read_input_token_cost: 2e-7
+    - cache_creation_input_token_cost_per_hour: 0.0000045
+      cache_read_input_token_cost: 2e-7
       input_cost_per_token: 0.000002
       input_cost_per_token_batches: 0.000001
       output_cost_per_token: 0.000012
@@ -46,6 +47,7 @@ params:
 provisioning: serverless
 sources:
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-1-pro
+    - https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview
 status: preview
 supportedModes:
     - chat

providers/google-vertex/google/gemma3n.yaml

@@ -3,18 +3,22 @@ features:
 limits:
     context_window: 32000
     max_input_tokens: 32000
+    max_output_tokens: 32000
+    max_tokens: 32000
 modalities:
     input:
         - text
         - image
         - audio
+        - video
     output:
         - text
 mode: chat
 model: google/gemma3n
 provisioning: provisioned
 sources:
     - https://ai.google.dev/gemma/docs/gemma-3n
+    - https://ai.google.dev/gemma/docs/gemma-3n/model_card
     - https://docs.cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-gemma
 status: active
 supportedModes:

providers/google-vertex/google/gemma4.yaml

@@ -16,6 +16,8 @@ modalities:
     input:
         - text
         - image
+        - audio
+        - video
         - pdf
     output:
         - text

providers/google-vertex/google/language-v1-classify-text-v1.yaml

@@ -1,5 +1,5 @@
 costs:
-    - input_cost_per_character: 0.000002
+    - input_cost_per_character: 2e-6
       region: "*"
 modalities:
     input:
@@ -13,7 +13,7 @@ removeParams:
     - max_tokens
     - temperature
     - top_p
-    - "n"
+    - n
     - stop
     - stream
     - tool_choice

providers/google-vertex/google/lyria-3-pro-preview.yaml

@@ -7,6 +7,7 @@ modalities:
         - image
     output:
         - audio
+        - text
 mode: unknown # correct mode is music, but that is not supported yet
 model: lyria-3-pro-preview
 provisioning: serverless