Management API Reference for gemini-chat-model

Chat model for Google Gemini

Provided Capabilities (1)

Name	Dynamic	Other provider points
org.wildfly.ai.chatmodel	true	/subsystem=ai/openai-chat-model=* /subsystem=ai/ollama-chat-model=* /subsystem=ai/github-chat-model=* /subsystem=ai/mistral-ai-chat-model=*

Attributes (28)

allowed-code-execution - Enables the model to generate and run Python code and learn iteratively from the results until it arrives at a final output.

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Enables the model to generate and run Python code and learn iteratively from the results until it arrives at a final output.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

api-key - API key to authenticate to a Google Gemini chat model.

Type	STRING
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "API key to authenticate to a Google Gemini chat model.", "expressions-allowed" : true, "required" : true, "nillable" : false, "min-length" : 1, "max-length" : 2147483647, "stability" : "default", "access-constraints" : {"sensitive" : {"credential" : {"type" : "core"}}}, "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

civic-integrity - Defines the threshold to filter out civic integrity content.

Type	STRING
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Allowed Values	BLOCK_NONE BLOCK_ONLY_HIGH BLOCK_LOW_AND_ABOVE BLOCK_MEDIUM_AND_ABOVE HARM_BLOCK_THRESHOLD_UNSPECIFIED
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "Defines the threshold to filter out civic integrity content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "min-length" : 1, "max-length" : 2147483647, "allowed" : [ "HARM_BLOCK_THRESHOLD_UNSPECIFIED", "BLOCK_LOW_AND_ABOVE", "BLOCK_MEDIUM_AND_ABOVE", "BLOCK_ONLY_HIGH", "BLOCK_NONE" ], "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

connect-timeout - Timeout for the Google Gemini chat model.

Type	LONG
Default Value	0
Unit	MILLISECONDS
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "LONG" }, "description" : "Timeout for the Google Gemini chat model.", "expressions-allowed" : true, "required" : false, "nillable" : true, "default" : 0, "unit" : "MILLISECONDS", "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

dangerous-content - Defines the threshold to filter out dangerous content.

Type	STRING
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Allowed Values	BLOCK_NONE BLOCK_ONLY_HIGH BLOCK_LOW_AND_ABOVE BLOCK_MEDIUM_AND_ABOVE HARM_BLOCK_THRESHOLD_UNSPECIFIED
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "Defines the threshold to filter out dangerous content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "min-length" : 1, "max-length" : 2147483647, "allowed" : [ "HARM_BLOCK_THRESHOLD_UNSPECIFIED", "BLOCK_LOW_AND_ABOVE", "BLOCK_MEDIUM_AND_ABOVE", "BLOCK_ONLY_HIGH", "BLOCK_NONE" ], "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

enable-enhanced-civic-answers - Enables enhanced civic answers. It may not be available for all models.

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Enables enhanced civic answers. It may not be available for all models.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

frequency-penalty - Positive values penalize tokens that repeatedly appear in the generated text, decreasing the probability of repeating content.

Type	DOUBLE
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "DOUBLE" }, "description" : "Positive values penalize tokens that repeatedly appear in the generated text, decreasing the probability of repeating content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

harassment - Defines the threshold to filter out harassment content.

Type	STRING
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Allowed Values	BLOCK_NONE BLOCK_ONLY_HIGH BLOCK_LOW_AND_ABOVE BLOCK_MEDIUM_AND_ABOVE HARM_BLOCK_THRESHOLD_UNSPECIFIED
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "Defines the threshold to filter out harassment content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "min-length" : 1, "max-length" : 2147483647, "allowed" : [ "HARM_BLOCK_THRESHOLD_UNSPECIFIED", "BLOCK_LOW_AND_ABOVE", "BLOCK_MEDIUM_AND_ABOVE", "BLOCK_ONLY_HIGH", "BLOCK_NONE" ], "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

hate-speech - Defines the threshold to filter out hate speech content.

Type	STRING
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Allowed Values	BLOCK_NONE BLOCK_ONLY_HIGH BLOCK_LOW_AND_ABOVE BLOCK_MEDIUM_AND_ABOVE HARM_BLOCK_THRESHOLD_UNSPECIFIED
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "Defines the threshold to filter out hate speech content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "min-length" : 1, "max-length" : 2147483647, "allowed" : [ "HARM_BLOCK_THRESHOLD_UNSPECIFIED", "BLOCK_LOW_AND_ABOVE", "BLOCK_MEDIUM_AND_ABOVE", "BLOCK_ONLY_HIGH", "BLOCK_NONE" ], "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

include-code-execution-output - Enables providing the code and the result in the output.

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Enables providing the code and the result in the output.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

include-thoughts - Enable the return of the thought summaries of the model.

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Enable the return of the thought summaries of the model.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

log-probs - Returns the log probabilities of the top candidate tokens at each generation step. The model's chosen token might not be the same as the top candidate token at each step. Specify the number of candidates to return by using an integer value in the range of 1-20.

Type	INT
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "INT" }, "description" : "Returns the log probabilities of the top candidate tokens at each generation step. The model's chosen token might not be the same as the top candidate token at each step. Specify the number of candidates to return by using an integer value in the range of 1-20. ", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

log-requests - Enabling the tracing of requests to Google Gemini

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Enabling the tracing of requests to Google Gemini", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

log-responses - Enabling the tracing of responses to Google Gemini

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Enabling the tracing of responses to Google Gemini", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

max-output-token - The number of token returned by the Google Gemini chat model.

Type	INT
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "INT" }, "description" : "The number of token returned by the Google Gemini chat model.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

model-name - Name of the model served by Google Gemini.

Type	STRING
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "Name of the model served by Google Gemini.", "expressions-allowed" : true, "required" : true, "nillable" : false, "min-length" : 1, "max-length" : 2147483647, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

presence-penalty - Presence penalty applied to the next token's logprobs if the token has already been seen in the response. This penalty is binary on/off and not dependant on the number of times the token is used (after the first). Use frequency-penalty for a penalty that increases with each use.

Type	DOUBLE
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "DOUBLE" }, "description" : "Presence penalty applied to the next token's logprobs if the token has already been seen in the response. This penalty is binary on/off and not dependant on the number of times the token is used (after the first). Use frequency-penalty for a penalty that increases with each use.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

response-format - The format of the response from Google Gemini.

Type	STRING
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Allowed Values	JSON TEXT
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "The format of the response from Google Gemini.", "expressions-allowed" : true, "required" : false, "nillable" : true, "allowed" : [ "JSON", "TEXT" ], "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

response-log-probs - If true, returns the log probabilities of the tokens that were chosen by the model at each step.

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "If true, returns the log probabilities of the tokens that were chosen by the model at each step.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

return-thinking - Controls whether to return thinking/reasoning text (if available).

Type	BOOLEAN
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Controls whether to return thinking/reasoning text (if available).", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

seed - When seed is fixed to a specific value, the model makes a best effort to provide the same response for repeated requests. Deterministic output isn't guaranteed.

Type	INT
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "INT" }, "description" : "When seed is fixed to a specific value, the model makes a best effort to provide the same response for repeated requests. Deterministic output isn't guaranteed.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

sexually-explicit - Defines the threshold to filter out sexually explicit content.

Type	STRING
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Allowed Values	BLOCK_NONE BLOCK_ONLY_HIGH BLOCK_LOW_AND_ABOVE BLOCK_MEDIUM_AND_ABOVE HARM_BLOCK_THRESHOLD_UNSPECIFIED
Stability Level	default
Raw DMR model	{ "type" : { "TYPE_MODEL_VALUE" : "STRING" }, "description" : "Defines the threshold to filter out sexually explicit content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "min-length" : 1, "max-length" : 2147483647, "allowed" : [ "HARM_BLOCK_THRESHOLD_UNSPECIFIED", "BLOCK_LOW_AND_ABOVE", "BLOCK_MEDIUM_AND_ABOVE", "BLOCK_ONLY_HIGH", "BLOCK_NONE" ], "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }

stop-sequences - List of stop sequences to tell the model to stop generating content.

Type	LIST
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	all-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "LIST" }, "description" : "List of stop sequences to tell the model to stop generating content.", "expressions-allowed" : true, "required" : false, "nillable" : true, "min-length" : 0, "max-length" : 2147483647, "stability" : "default", "value-type" : { "TYPE_MODEL_VALUE" : "STRING" }, "access-type" : "read-write", "storage" : "configuration", "restart-required" : "all-services", "capability-reference" : null }`

streaming - Whether to create a token streaming chat language model or not.

Type	BOOLEAN
Default Value	false
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "BOOLEAN" }, "description" : "Whether to create a token streaming chat language model or not.", "expressions-allowed" : true, "required" : false, "nillable" : true, "default" : false, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

temperature - Temperature of the Google Gemini chat model.

Type	DOUBLE
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "DOUBLE" }, "description" : "Temperature of the Google Gemini chat model.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

thinking-budget - Parameter to guide the model on the number of thinking tokens to use when generating a response.

Type	INT
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "INT" }, "description" : "Parameter to guide the model on the number of thinking tokens to use when generating a response.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

top-k - TopK of the Google Gemini chat model.

Type	INT
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "INT" }, "description" : "TopK of the Google Gemini chat model.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

top-p - TopP of the Google Gemini chat model.

Type	DOUBLE
Is the attribute nillable?	true
Does the attribute allow expression?	true
Where is the attribute stored?	configuration
Access Type	read-write
Is restarted required?	no-services
Stability Level	default
Raw DMR model	`{ "type" : { "TYPE_MODEL_VALUE" : "DOUBLE" }, "description" : "TopP of the Google Gemini chat model.", "expressions-allowed" : true, "required" : false, "nillable" : true, "stability" : "default", "access-type" : "read-write", "storage" : "configuration", "restart-required" : "no-services", "capability-reference" : null }`

Operations (3)

add - This operation adds a Google Gemini chat model.

Request Parameter	Type	Required	Expressions Allowed	Default value	Description
allowed-code-execution	BOOLEAN	false	true		Enables the model to generate and run Python code and learn iteratively from the results until it arrives at a final output.
api-key	STRING	true	true		API key to authenticate to a Google Gemini chat model.
civic-integrity	STRING	false	true		Defines the threshold to filter out civic integrity content.
connect-timeout	LONG	false	true	0	Timeout for the Google Gemini chat model.
dangerous-content	STRING	false	true		Defines the threshold to filter out dangerous content.
enable-enhanced-civic-answers	BOOLEAN	false	true		Enables enhanced civic answers. It may not be available for all models.
frequency-penalty	DOUBLE	false	true		Positive values penalize tokens that repeatedly appear in the generated text, decreasing the probability of repeating content.
harassment	STRING	false	true		Defines the threshold to filter out harassment content.
hate-speech	STRING	false	true		Defines the threshold to filter out hate speech content.
include-code-execution-output	BOOLEAN	false	true		Enables providing the code and the result in the output.
include-thoughts	BOOLEAN	false	true		Enable the return of the thought summaries of the model.
log-probs	INT	false	true		Returns the log probabilities of the top candidate tokens at each generation step. The model's chosen token might not be the same as the top candidate token at each step. Specify the number of candidates to return by using an integer value in the range of 1-20.
log-requests	BOOLEAN	false	true		Enabling the tracing of requests to Google Gemini
log-responses	BOOLEAN	false	true		Enabling the tracing of responses to Google Gemini
max-output-token	INT	false	true		The number of token returned by the Google Gemini chat model.
model-name	STRING	true	true		Name of the model served by Google Gemini.
presence-penalty	DOUBLE	false	true		Presence penalty applied to the next token's logprobs if the token has already been seen in the response. This penalty is binary on/off and not dependant on the number of times the token is used (after the first). Use frequency-penalty for a penalty that increases with each use.
response-format	STRING	false	true		The format of the response from Google Gemini.
response-log-probs	BOOLEAN	false	true		If true, returns the log probabilities of the tokens that were chosen by the model at each step.
return-thinking	BOOLEAN	false	true		Controls whether to return thinking/reasoning text (if available).
seed	INT	false	true		When seed is fixed to a specific value, the model makes a best effort to provide the same response for repeated requests. Deterministic output isn't guaranteed.
sexually-explicit	STRING	false	true		Defines the threshold to filter out sexually explicit content.
stop-sequences	LIST	false	true		List of stop sequences to tell the model to stop generating content.
streaming	BOOLEAN	false	true	false	Whether to create a token streaming chat language model or not.
temperature	DOUBLE	false	true		Temperature of the Google Gemini chat model.
thinking-budget	INT	false	true		Parameter to guide the model on the number of thinking tokens to use when generating a response.
top-k	INT	false	true		TopK of the Google Gemini chat model.
top-p	DOUBLE	false	true		TopP of the Google Gemini chat model.

Raw DMR model


{
    "operation-name" : "add",
    "description" : "This operation adds a Google Gemini chat model.",
    "request-properties" : {
        "allowed-code-execution" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Enables the model to generate and run Python code and learn iteratively from the results until it arrives at a final output.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "api-key" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "API key to authenticate to a Google Gemini chat model.",
            "expressions-allowed" : true,
            "required" : true,
            "nillable" : false,
            "min-length" : 1,
            "max-length" : 2147483647,
            "stability" : "default"
        },
        "civic-integrity" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "Defines the threshold to filter out civic integrity content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "min-length" : 1,
            "max-length" : 2147483647,
            "allowed" : [
                "HARM_BLOCK_THRESHOLD_UNSPECIFIED",
                "BLOCK_LOW_AND_ABOVE",
                "BLOCK_MEDIUM_AND_ABOVE",
                "BLOCK_ONLY_HIGH",
                "BLOCK_NONE"
            ],
            "stability" : "default"
        },
        "connect-timeout" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "LONG"
            },
            "description" : "Timeout for the Google Gemini chat model.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "default" : 0,
            "unit" : "MILLISECONDS",
            "stability" : "default"
        },
        "dangerous-content" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "Defines the threshold to filter out dangerous content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "min-length" : 1,
            "max-length" : 2147483647,
            "allowed" : [
                "HARM_BLOCK_THRESHOLD_UNSPECIFIED",
                "BLOCK_LOW_AND_ABOVE",
                "BLOCK_MEDIUM_AND_ABOVE",
                "BLOCK_ONLY_HIGH",
                "BLOCK_NONE"
            ],
            "stability" : "default"
        },
        "enable-enhanced-civic-answers" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Enables enhanced civic answers. It may not be available for all models.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "frequency-penalty" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "DOUBLE"
            },
            "description" : "Positive values penalize tokens that repeatedly appear in the generated text, decreasing the probability of repeating content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "harassment" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "Defines the threshold to filter out harassment content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "min-length" : 1,
            "max-length" : 2147483647,
            "allowed" : [
                "HARM_BLOCK_THRESHOLD_UNSPECIFIED",
                "BLOCK_LOW_AND_ABOVE",
                "BLOCK_MEDIUM_AND_ABOVE",
                "BLOCK_ONLY_HIGH",
                "BLOCK_NONE"
            ],
            "stability" : "default"
        },
        "hate-speech" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "Defines the threshold to filter out hate speech content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "min-length" : 1,
            "max-length" : 2147483647,
            "allowed" : [
                "HARM_BLOCK_THRESHOLD_UNSPECIFIED",
                "BLOCK_LOW_AND_ABOVE",
                "BLOCK_MEDIUM_AND_ABOVE",
                "BLOCK_ONLY_HIGH",
                "BLOCK_NONE"
            ],
            "stability" : "default"
        },
        "include-code-execution-output" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Enables providing the code and the result in the output.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "include-thoughts" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Enable the return of the thought summaries of the model.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "log-probs" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "INT"
            },
            "description" : "Returns the log probabilities of the top candidate tokens at each generation step. The model's chosen token might not be the same as the top candidate token at each step. Specify the number of candidates to return by using an integer value in the range of 1-20. ",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "log-requests" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Enabling the tracing of requests to Google Gemini",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "log-responses" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Enabling the tracing of responses to Google Gemini",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "max-output-token" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "INT"
            },
            "description" : "The number of token returned by the Google Gemini chat model.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "model-name" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "Name of the model served by Google Gemini.",
            "expressions-allowed" : true,
            "required" : true,
            "nillable" : false,
            "min-length" : 1,
            "max-length" : 2147483647,
            "stability" : "default"
        },
        "presence-penalty" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "DOUBLE"
            },
            "description" : "Presence penalty applied to the next token's logprobs if the token has already been seen in the response. This penalty is binary on/off and not dependant on the number of times the token is used (after the first). Use frequency-penalty for a penalty that increases with each use.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "response-format" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "The format of the response from Google Gemini.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "allowed" : [
                "JSON",
                "TEXT"
            ],
            "stability" : "default"
        },
        "response-log-probs" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "If true, returns the log probabilities of the tokens that were chosen by the model at each step.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "return-thinking" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Controls whether to return thinking/reasoning text (if available).",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "seed" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "INT"
            },
            "description" : "When seed is fixed to a specific value, the model makes a best effort to provide the same response for repeated requests. Deterministic output isn't guaranteed.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "sexually-explicit" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            },
            "description" : "Defines the threshold to filter out sexually explicit content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "min-length" : 1,
            "max-length" : 2147483647,
            "allowed" : [
                "HARM_BLOCK_THRESHOLD_UNSPECIFIED",
                "BLOCK_LOW_AND_ABOVE",
                "BLOCK_MEDIUM_AND_ABOVE",
                "BLOCK_ONLY_HIGH",
                "BLOCK_NONE"
            ],
            "stability" : "default"
        },
        "stop-sequences" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "LIST"
            },
            "description" : "List of stop sequences to tell the model to stop generating content.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "min-length" : 0,
            "max-length" : 2147483647,
            "stability" : "default",
            "value-type" : {
                "TYPE_MODEL_VALUE" : "STRING"
            }
        },
        "streaming" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "BOOLEAN"
            },
            "description" : "Whether to create a token streaming chat language model or not.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "default" : false,
            "stability" : "default"
        },
        "temperature" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "DOUBLE"
            },
            "description" : "Temperature of the Google Gemini chat model.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "thinking-budget" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "INT"
            },
            "description" : "Parameter to guide the model on the number of thinking tokens to use when generating a response.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "top-k" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "INT"
            },
            "description" : "TopK of the Google Gemini chat model.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        },
        "top-p" : {
            "type" : {
                "TYPE_MODEL_VALUE" : "DOUBLE"
            },
            "description" : "TopP of the Google Gemini chat model.",
            "expressions-allowed" : true,
            "required" : false,
            "nillable" : true,
            "stability" : "default"
        }
    },
    "reply-properties" : {},
    "stability" : "default",
    "read-only" : false,
    "runtime-only" : false
}

chat - Simple operation to send a user message to the LLM.

Request Parameter	Type	Required	Expressions Allowed	Default value	Description
user-message	STRING	true	true		The user message sent to the LLM to test it.

Reply properties

Type: STRING

Raw DMR model


{
    "operation-name" : "chat",
    "description" : "Simple operation to send a user message to the LLM.",
    "request-properties" : {"user-message" : {
        "type" : {
            "TYPE_MODEL_VALUE" : "STRING"
        },
        "description" : "The user message sent to the LLM to test it.",
        "expressions-allowed" : true,
        "required" : true,
        "nillable" : false,
        "min-length" : 1,
        "max-length" : 2147483647,
        "stability" : "default"
    }},
    "reply-properties" : {"type" : {
        "TYPE_MODEL_VALUE" : "STRING"
    }},
    "access-constraints" : {"sensitive" : {"read-whole-config" : {"type" : "core"}}},
    "stability" : "default",
    "read-only" : true,
    "runtime-only" : true
}

remove - This operation removes a Google Gemini chat model.

Raw DMR model


{
    "operation-name" : "remove",
    "description" : "This operation removes a Google Gemini chat model.",
    "request-properties" : {},
    "reply-properties" : {},
    "stability" : "default",
    "read-only" : false,
    "restart-required" : "resource-services",
    "runtime-only" : false
}

Galleon Features

Galleon features are to be used when creating Galleon feature-packs.

subsystem.ai.gemini-chat-model

        
<feature spec="subsystem.ai.gemini-chat-model">
  <param name="gemini-chat-model" value="{resource name}"/>
  <param name="allowed-code-execution" value="{value}"/>
  <param name="api-key" value="{value}"/>
  <param name="civic-integrity" value="{value}"/>
  <param name="connect-timeout" value="{value}"/>
  <param name="dangerous-content" value="{value}"/>
  <param name="enable-enhanced-civic-answers" value="{value}"/>
  <param name="frequency-penalty" value="{value}"/>
  <param name="harassment" value="{value}"/>
  <param name="hate-speech" value="{value}"/>
  <param name="include-code-execution-output" value="{value}"/>
  <param name="include-thoughts" value="{value}"/>
  <param name="log-probs" value="{value}"/>
  <param name="log-requests" value="{value}"/>
  <param name="log-responses" value="{value}"/>
  <param name="max-output-token" value="{value}"/>
  <param name="model-name" value="{value}"/>
  <param name="presence-penalty" value="{value}"/>
  <param name="response-format" value="{value}"/>
  <param name="response-log-probs" value="{value}"/>
  <param name="return-thinking" value="{value}"/>
  <param name="seed" value="{value}"/>
  <param name="sexually-explicit" value="{value}"/>
  <param name="stop-sequences" value="{value}"/>
  <param name="streaming" value="{value}"/>
  <param name="temperature" value="{value}"/>
  <param name="thinking-budget" value="{value}"/>
  <param name="top-k" value="{value}"/>
  <param name="top-p" value="{value}"/>
<feature/>