W-20781836 Create LLM table and link from agent-network reference

modules/ROOT/pages/af-agent-networks.adoc

@@ -60,6 +60,39 @@ These servers will be published to Exchange.
 The asset might also be published to Exchange.
 . An asset that has been added to this agent network project from Exchange.
 
+[[llm-support]]
+== Large Language Models
+
+Agent network brokers support these LLMs.
+
+* Azure OpenAI
+* OpenAI API
+* Gemini API
+
+The following table details the requirements and recommended models. 
+
+|===
+|*Model Provider* |*Required Endpoint* |*Required Capabilities* |*Suggested Models*
+
+|OpenAI |`/responses` a|
+* Reasoning 
+* Native structured output 
+* Function and Custom tool calling 
+a|
+* For lower latency: GPT-5-mini 
+* For complex reasoning: Evaluate models
+
+|Gemini |`/generateContent` (Native API) a|
+* Native Thinking (via thinkingConfig) 
+* Native structured output (responseSchema) 
+* Function and custom tool calling 
+a|
+* For lower latency: Gemini 2.5 Flash, Gemini 2.5 Flash-Lite 
+* For complex reasoning: Gemini 3 Pro (Deep Think capabilities)
+|=== 
+
+Agent network supports text-based prompts and responses. Image and binary message types aren't supported.
+
 == A2A Protocol 
 
 The Agent2Agent (A2A) Protocol governs agent-to-agent communication. This protocol powers orchestration, observability, and governance features in agent networks. MuleSoft supports v0.3.0 of the A2A Protocol Specification.
@@ -89,28 +122,8 @@ Consider a network with a client and two brokers (1 and 2).
 
 For more information, see https://a2a-protocol.org/v0.3.0/topics/life-of-a-task/#group-related-interactions[Life of a Task - Group Related Interactions].
 
-=== Non-Supported Features
-
-Agent network doesn't support streaming with Server-Sent Events (SSE).
-
-== Considerations
-
-Agent networks have these considerations.
+NOTE: Agent network doesn't support streaming with Server-Sent Events (SSE).
 
-* An agent network project supports only one domain (one YAML file per project). For multiple domains, create a separate project for each.
-* These authentication types are supported.
-** Basic
-** Client Credentials
-** Anypoint Client Credentials
-* Text-based prompts and responses are supported. Image and binary message/artifact types aren't supported.
-* All LLM models in your agent network project must support structured responses in JSON format. For example, a `gpt4.1` model is supported; a `gpt4` model isn't.
-* These LLMs are supported.
-** Azure OpenAI
-** OpenAI Direct API
-* When you add governance policies to your agent network project YAML file, those policies persist between deployments. However, if you add governance policies at run-time, those policies don't persist. The next time you deploy your agent network, the governance policies used are the ones in the `agent-network.yaml` file.
-* These deployment targets are supported.
-** CloudHub 2.0
-** Runtime Fabric (limited availability)
 
 == See Also
 

modules/ROOT/pages/af-project-files.adoc

@@ -231,14 +231,7 @@ The `spec` element has these properties.
 
 The value of this section is a reference to one of the LLMs defined in Anypoint Exchange or in the `llmProviders` section of `agent-network.yaml`. Because it's a reference, you can choose to share the same LLM across all the brokers in your agent network. Or, you can have different brokers use different LLMs to better suit their tasks.
 
-Agent network brokers support OpenAI models. The model you use must support: 
-
-* The `/responses` endpoint
-* Reasoning
-* Native structured output
-* Function and Custom tool calling
-
-GPT-5-mini and GPT-5.1 meet these requirements. GPT-5-mini and GPT-5.1 meet these requirements. Evaluate the different models to find the right balance for your needs. For lower latency, consider smaller models like GPT-5-mini.
+For more information on supported LLMs, see xref:af-agent-networks.adoc#llm-support[].
 
 The `llm` element has these properties.
 
@@ -264,7 +257,17 @@ a|
 |`configuration.openAI.topP` |Nucleus sampling parameter. Requires GPT-5.1 and `reasoningEffort` set to `NONE`. |Number |Any number |No
 |`configuration.openAI.topLogprobs` |Number of most likely tokens to return at each position. Requires GPT-5.1 and `reasoningEffort` set to `NONE`. |Integer |Any integer |No
 |`configuration.openAI.maxOutputTokens` |Maximum number of tokens to generate. |Integer |Any integer |No
-|===
+|`configuration.gemini` |OpenAI specific configuration settings. |Object |Object with OpenAI settings |No
+|`configuration.gemini.thinkingConfig.thinkingBudget` |Sets a token budget for the reasoning phase. (Applies only to Gemini 2.5 series.) |Integer a|
+* `0` (Disabled)
+* `1024` to `32768` |No
+|`configuration.gemini.thinkingConfig.thinkingLevel` |Controls the depth of the reasoning process. (Applies only to Gemini 3 series.) |String a|
+* `High` (Default)
+* `Low` |No
+|`configuration.gemini.temperature` |Controls randomness. |Number |For Gemini 3 and Gemini 2.5, Google recommends keeping this at `1.0` (default) to avoid breaking the reasoning chain. |No
+|`configuration.gemini.topP` |Nucleus sampling parameter. |Number |Any number |No
+|`configuration.gemini.responseLogprobs` |Whether to return log probabilities. |Boolean |true or false |No
+|=== 
 
 [[instructions-section]]
 ==== Instructions Section