diff --git a/fern/apis/api/openapi.json b/fern/apis/api/openapi.json
index 5a658c25b..33714dd43 100644
--- a/fern/apis/api/openapi.json
+++ b/fern/apis/api/openapi.json
@@ -13052,6 +13052,7 @@
             "type": "string",
             "description": "This is the model that will be used for the transcription.",
             "enum": [
+              "gemini-3.8-flash",
               "gemini-3.5-flash",
               "gemini-3.1-flash-lite",
               "gemini-3-flash-preview",
@@ -16104,6 +16105,7 @@
             "type": "string",
             "description": "This is the model that will be used for the transcription.",
             "enum": [
+              "gemini-3.8-flash",
               "gemini-3.5-flash",
               "gemini-3.1-flash-lite",
               "gemini-3-flash-preview",
@@ -19134,6 +19136,7 @@
             "type": "string",
             "description": "The model to use for the knowledge base",
             "enum": [
+              "gemini-3.8-flash",
               "gemini-3.5-flash",
               "gemini-3.1-flash-lite",
               "gemini-3-flash-preview",
@@ -19909,7 +19912,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -20100,7 +20103,20 @@
               "claude-sonnet-4-5-20250929",
               "claude-sonnet-4-6",
               "claude-sonnet-5",
-              "claude-haiku-4-5-20251001"
+              "claude-haiku-4-5-20251001",
+              "claude-haiku-5-5"
+            ]
+          },
+          "reasoningEffort": {
+            "type": "string",
+            "description": "Reasoning effort for claude-haiku-5-5. `none` turns thinking off. `low` to `max`\nuse adaptive thinking at that effort; higher effort can add latency before the\nmodel speaks. Unset means `none`. Rejected for every other model.",
+            "enum": [
+              "none",
+              "low",
+              "medium",
+              "high",
+              "xhigh",
+              "max"
             ]
           },
           "provider": {
@@ -20126,7 +20142,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -20303,6 +20319,7 @@
               "claude-sonnet-4-5-20250929",
               "claude-sonnet-4-6",
               "claude-haiku-4-5-20251001",
+              "claude-haiku-5-5",
               "global.anthropic.claude-haiku-4-5-20251001-v1:0"
             ]
           },
@@ -20324,6 +20341,7 @@
               "claude-sonnet-4-5-20250929",
               "claude-sonnet-4-6",
               "claude-haiku-4-5-20251001",
+              "claude-haiku-5-5",
               "global.anthropic.claude-haiku-4-5-20251001-v1:0"
             ],
             "maxItems": 1,
@@ -20347,10 +20365,23 @@
                 "claude-sonnet-4-5-20250929",
                 "claude-sonnet-4-6",
                 "claude-haiku-4-5-20251001",
+                "claude-haiku-5-5",
                 "global.anthropic.claude-haiku-4-5-20251001-v1:0"
               ]
             }
           },
+          "reasoningEffort": {
+            "type": "string",
+            "description": "Reasoning effort for claude-haiku-5-5. `none` turns thinking off. `low` to `max`\nuse adaptive thinking at that effort; higher effort can add latency before the\nmodel speaks. Unset means `none`. Rejected for every other model.",
+            "enum": [
+              "none",
+              "low",
+              "medium",
+              "high",
+              "xhigh",
+              "max"
+            ]
+          },
           "thinking": {
             "description": "Optional configuration for Anthropic's thinking feature.\nOnly applicable for claude-3-7-sonnet-20250219 model.\nIf provided, maxTokens must be greater than thinking.budgetTokens.",
             "allOf": [
@@ -20367,7 +20398,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -20541,7 +20572,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -20745,7 +20776,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -20916,7 +20947,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -21092,7 +21123,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -21317,6 +21348,7 @@
             "type": "string",
             "description": "This is the Google model that will be used.",
             "enum": [
+              "gemini-3.8-flash",
               "gemini-3.5-flash",
               "gemini-3.1-flash-lite",
               "gemini-3-flash-preview",
@@ -21358,7 +21390,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -21544,7 +21576,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -21717,7 +21749,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -21890,7 +21922,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -22584,7 +22616,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -22754,7 +22786,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -22924,7 +22956,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -23094,7 +23126,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -23144,7 +23176,7 @@
         "type": "object",
         "properties": {
           "messages": {
-            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. Message contents support LiquidJS templating with the following variables:\n- `{{transcript}}` or `{{messages}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\n`{{messages}}` is the full message history including tool calls; `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
+            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. For text extraction, message contents support LiquidJS templating with the following variables:\n- `{{transcript}}`, `{{messages}}`, or `{{messagesWithConfidence}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results, but omits caller transcription-confidence fields. For text extraction, `{{messagesWithConfidence}}` includes the original message history with transcription confidence for outputs that explicitly need it. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data. `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
             "type": "array",
             "items": {
               "$ref": "#/components/schemas/OpenAIMessage"
@@ -23337,7 +23369,7 @@
         "type": "object",
         "properties": {
           "messages": {
-            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. Message contents support LiquidJS templating with the following variables:\n- `{{transcript}}` or `{{messages}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\n`{{messages}}` is the full message history including tool calls; `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
+            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. For text extraction, message contents support LiquidJS templating with the following variables:\n- `{{transcript}}`, `{{messages}}`, or `{{messagesWithConfidence}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results, but omits caller transcription-confidence fields. For text extraction, `{{messagesWithConfidence}}` includes the original message history with transcription confidence for outputs that explicitly need it. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data. `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
             "type": "array",
             "items": {
               "$ref": "#/components/schemas/OpenAIMessage"
@@ -23369,7 +23401,8 @@
               "claude-sonnet-4-5-20250929",
               "claude-sonnet-4-6",
               "claude-sonnet-5",
-              "claude-haiku-4-5-20251001"
+              "claude-haiku-4-5-20251001",
+              "claude-haiku-5-5"
             ]
           },
           "thinking": {
@@ -23402,7 +23435,7 @@
         "type": "object",
         "properties": {
           "messages": {
-            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. Message contents support LiquidJS templating with the following variables:\n- `{{transcript}}` or `{{messages}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\n`{{messages}}` is the full message history including tool calls; `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
+            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. For text extraction, message contents support LiquidJS templating with the following variables:\n- `{{transcript}}`, `{{messages}}`, or `{{messagesWithConfidence}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results, but omits caller transcription-confidence fields. For text extraction, `{{messagesWithConfidence}}` includes the original message history with transcription confidence for outputs that explicitly need it. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data. `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
             "type": "array",
             "items": {
               "$ref": "#/components/schemas/OpenAIMessage"
@@ -23434,6 +23467,7 @@
               "claude-sonnet-4-5-20250929",
               "claude-sonnet-4-6",
               "claude-haiku-4-5-20251001",
+              "claude-haiku-5-5",
               "global.anthropic.claude-haiku-4-5-20251001-v1:0"
             ]
           },
@@ -23467,7 +23501,7 @@
         "type": "object",
         "properties": {
           "messages": {
-            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. Message contents support LiquidJS templating with the following variables:\n- `{{transcript}}` or `{{messages}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\n`{{messages}}` is the full message history including tool calls; `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
+            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. For text extraction, message contents support LiquidJS templating with the following variables:\n- `{{transcript}}`, `{{messages}}`, or `{{messagesWithConfidence}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results, but omits caller transcription-confidence fields. For text extraction, `{{messagesWithConfidence}}` includes the original message history with transcription confidence for outputs that explicitly need it. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data. `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
             "type": "array",
             "items": {
               "$ref": "#/components/schemas/OpenAIMessage"
@@ -23485,6 +23519,7 @@
             "description": "This is the name of the model. Ex. cognitivecomputations/dolphin-mixtral-8x7b",
             "maxLength": 100,
             "enum": [
+              "gemini-3.8-flash",
               "gemini-3.5-flash",
               "gemini-3.1-flash-lite",
               "gemini-3-flash-preview",
@@ -23526,7 +23561,7 @@
         "type": "object",
         "properties": {
           "messages": {
-            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. Message contents support LiquidJS templating with the following variables:\n- `{{transcript}}` or `{{messages}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\n`{{messages}}` is the full message history including tool calls; `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
+            "description": "These are the messages used to customize the prompt used for structured output extraction.\n\nWhen provided, these messages replace the default prompts. For text extraction, message contents support LiquidJS templating with the following variables:\n- `{{transcript}}`, `{{messages}}`, or `{{messagesWithConfidence}}` to reference the conversation (one is required)\n- `{{structuredOutput.name}}`, `{{structuredOutput.description}}`, or `{{structuredOutput.schema}}` to reference the structured output definition (one is required)\n- `{{systemPrompt}}`, `{{callEndedReason}}`, `{{duration}}`, `{{startedAt}}`, `{{endedAt}}`, and any `assistantOverrides.variableValues`\n\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results, but omits caller transcription-confidence fields. For text extraction, `{{messagesWithConfidence}}` includes the original message history with transcription confidence for outputs that explicitly need it. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data. `{{transcript}}` is the spoken text only, which uses significantly fewer tokens.\n\nIf not provided, default system and user prompts are used.",
             "type": "array",
             "items": {
               "$ref": "#/components/schemas/OpenAIMessage"
@@ -25525,7 +25560,7 @@
             "maxLength": 1000
           },
           "model": {
-            "description": "This is the model that will be used to extract the structured output.\n\nTo provide your own custom system and user prompts for structured output extraction, populate the messages array with your system and user messages. You can specify liquid templating in your system and user messages.\nBetween the system or user messages, you must reference either 'transcript' or 'messages' with the `{{}}` syntax to access the conversation history.\nBetween the system or user messages, you must reference a variation of the structured output with the `{{}}` syntax to access the structured output definition.\ni.e.:\n`{{structuredOutput}}`\n`{{structuredOutput.name}}`\n`{{structuredOutput.description}}`\n`{{structuredOutput.schema}}`\n\nIf model is not specified, GPT-4.1 will be used by default for extraction, utilizing default system and user prompts.\nIf messages or required fields are not specified, the default system and user prompts will be used.",
+            "description": "This is the model that will be used to extract the structured output.\n\nTo provide your own custom system and user prompts for structured output extraction, populate the messages array with your system and user messages. You can specify liquid templating in your system and user messages.\nFor text extraction, you must reference 'transcript', 'messages', or 'messagesWithConfidence' with the `{{}}` syntax to access the conversation history.\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results but omits caller transcription-confidence fields. Text extraction can use `{{messagesWithConfidence}}` when the output needs the original message history with transcription confidence. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data.\nBetween the system or user messages, you must reference a variation of the structured output with the `{{}}` syntax to access the structured output definition.\ni.e.:\n`{{structuredOutput}}`\n`{{structuredOutput.name}}`\n`{{structuredOutput.description}}`\n`{{structuredOutput.schema}}`\n\nIf model is not specified, GPT-4.1 will be used by default for extraction, utilizing default system and user prompts.\nIf messages or required fields are not specified, the default system and user prompts will be used.",
             "oneOf": [
               {
                 "$ref": "#/components/schemas/WorkflowOpenAIModel",
@@ -27090,7 +27125,7 @@
           },
           "maxTokens": {
             "type": "number",
-            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna no cap is applied unless you set one, because reasoning uses output tokens and a small cap can leave the reply empty.",
+            "description": "This is the max number of tokens that the assistant will be allowed to generate in each turn of the conversation. Default is 250. On gpt-6-luna and claude-haiku-5-5 no cap is applied unless you set one (claude-haiku-5-5 sends its 128,000-token maximum), because reasoning uses output tokens and a small cap can leave the reply empty.",
             "minimum": 50,
             "maximum": 10000
           },
@@ -27608,6 +27643,8 @@
             "enum": [
               "sonic-3.5",
               "sonic-3.5-2026-05-04",
+              "sonic-3.6",
+              "sonic-3.6-2026-08-27",
               "sonic-3",
               "sonic-3-2026-01-12",
               "sonic-3-2025-10-27",
@@ -27651,6 +27688,7 @@
               "ms",
               "nl",
               "no",
+              "or",
               "pa",
               "pl",
               "pt",
@@ -27664,6 +27702,7 @@
               "tl",
               "tr",
               "uk",
+              "ur",
               "vi",
               "zh"
             ],
@@ -30496,6 +30535,8 @@
             "enum": [
               "sonic-3.5",
               "sonic-3.5-2026-05-04",
+              "sonic-3.6",
+              "sonic-3.6-2026-08-27",
               "sonic-3",
               "sonic-3-2026-01-12",
               "sonic-3-2025-10-27",
@@ -30539,6 +30580,7 @@
               "ms",
               "nl",
               "no",
+              "or",
               "pa",
               "pl",
               "pt",
@@ -30552,6 +30594,7 @@
               "tl",
               "tr",
               "uk",
+              "ur",
               "vi",
               "zh"
             ],
@@ -59400,7 +59443,7 @@
             "maxLength": 1000
           },
           "model": {
-            "description": "This is the model that will be used to extract the structured output.\n\nTo provide your own custom system and user prompts for structured output extraction, populate the messages array with your system and user messages. You can specify liquid templating in your system and user messages.\nBetween the system or user messages, you must reference either 'transcript' or 'messages' with the `{{}}` syntax to access the conversation history.\nBetween the system or user messages, you must reference a variation of the structured output with the `{{}}` syntax to access the structured output definition.\ni.e.:\n`{{structuredOutput}}`\n`{{structuredOutput.name}}`\n`{{structuredOutput.description}}`\n`{{structuredOutput.schema}}`\n\nIf model is not specified, GPT-4.1 will be used by default for extraction, utilizing default system and user prompts.\nIf messages or required fields are not specified, the default system and user prompts will be used.",
+            "description": "This is the model that will be used to extract the structured output.\n\nTo provide your own custom system and user prompts for structured output extraction, populate the messages array with your system and user messages. You can specify liquid templating in your system and user messages.\nFor text extraction, you must reference 'transcript', 'messages', or 'messagesWithConfidence' with the `{{}}` syntax to access the conversation history.\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results but omits caller transcription-confidence fields. Text extraction can use `{{messagesWithConfidence}}` when the output needs the original message history with transcription confidence. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data.\nBetween the system or user messages, you must reference a variation of the structured output with the `{{}}` syntax to access the structured output definition.\ni.e.:\n`{{structuredOutput}}`\n`{{structuredOutput.name}}`\n`{{structuredOutput.description}}`\n`{{structuredOutput.schema}}`\n\nIf model is not specified, GPT-4.1 will be used by default for extraction, utilizing default system and user prompts.\nIf messages or required fields are not specified, the default system and user prompts will be used.",
             "oneOf": [
               {
                 "$ref": "#/components/schemas/WorkflowOpenAIModel",
@@ -59604,7 +59647,7 @@
             }
           },
           "model": {
-            "description": "This is the model that will be used to extract the structured output.\n\nTo provide your own custom system and user prompts for structured output extraction, populate the messages array with your system and user messages. You can specify liquid templating in your system and user messages.\nBetween the system or user messages, you must reference either 'transcript' or 'messages' with the `{{}}` syntax to access the conversation history.\nBetween the system or user messages, you must reference a variation of the structured output with the `{{}}` syntax to access the structured output definition.\ni.e.:\n`{{structuredOutput}}`\n`{{structuredOutput.name}}`\n`{{structuredOutput.description}}`\n`{{structuredOutput.schema}}`\n\nIf model is not specified, GPT-4.1 will be used by default for extraction, utilizing default system and user prompts.\nIf messages or required fields are not specified, the default system and user prompts will be used.",
+            "description": "This is the model that will be used to extract the structured output.\n\nTo provide your own custom system and user prompts for structured output extraction, populate the messages array with your system and user messages. You can specify liquid templating in your system and user messages.\nFor text extraction, you must reference 'transcript', 'messages', or 'messagesWithConfidence' with the `{{}}` syntax to access the conversation history.\nIn structured outputs, `{{messages}}` includes conversation text and tool calls/results but omits caller transcription-confidence fields. Text extraction can use `{{messagesWithConfidence}}` when the output needs the original message history with transcription confidence. These variables also apply in structured-output schema descriptions. Audio-targeted (`messages-with-audio`) extraction ignores custom messages. Stored call artifacts and API messages retain their confidence data.\nBetween the system or user messages, you must reference a variation of the structured output with the `{{}}` syntax to access the structured output definition.\ni.e.:\n`{{structuredOutput}}`\n`{{structuredOutput.name}}`\n`{{structuredOutput.description}}`\n`{{structuredOutput.schema}}`\n\nIf model is not specified, GPT-4.1 will be used by default for extraction, utilizing default system and user prompts.\nIf messages or required fields are not specified, the default system and user prompts will be used.",
             "oneOf": [
               {
                 "$ref": "#/components/schemas/WorkflowOpenAIModel",
@@ -64100,6 +64143,8 @@
               "call.hookTriggered",
               "call.hookSucceeded",
               "call.hookFailed",
+              "serverMessage.requestSucceeded",
+              "serverMessage.requestFailed",
               "call.statusReceived",
               "call.silenceTimeout",
               "call.microphoneTimeout",
@@ -64168,7 +64213,8 @@
               "pipeline.userSpeechStopped",
               "pipeline.endpointingTriggered",
               "pipeline.firstMessageStarted",
-              "pipeline.firstMessageCompleted"
+              "pipeline.firstMessageCompleted",
+              "pipeline.denoiserFailed"
             ]
           },
           "operation": {
@@ -65901,7 +65947,8 @@
               "claude-sonnet-4-5-20250929",
               "claude-sonnet-4-6",
               "claude-sonnet-5",
-              "claude-haiku-4-5-20251001"
+              "claude-haiku-4-5-20251001",
+              "claude-haiku-5-5"
             ]
           },
           "thinking": {
@@ -65954,6 +66001,7 @@
             "description": "This is the name of the model. Ex. gpt-4o",
             "maxLength": 100,
             "enum": [
+              "gemini-3.8-flash",
               "gemini-3.5-flash",
               "gemini-3.1-flash-lite",
               "gemini-3-flash-preview",