File tree Expand file tree Collapse file tree
Expand file tree Collapse file tree Original file line number Diff line number Diff line change @@ -1115,6 +1115,35 @@ public class ProviderConfig
11151115 /// </summary>
11161116 [ JsonPropertyName ( "azure" ) ]
11171117 public AzureOptions ? Azure { get ; set ; }
1118+
1119+ /// <summary>
1120+ /// Overrides the maximum number of output tokens the model can generate.
1121+ /// When set, takes precedence over the default limit resolved from the model's capability catalog entry.
1122+ /// </summary>
1123+ [ JsonPropertyName ( "maxOutputTokens" ) ]
1124+ public int ? MaxOutputTokens { get ; set ; }
1125+
1126+ /// <summary>
1127+ /// Overrides the maximum number of prompt/input tokens.
1128+ /// When set, takes precedence over the default limit resolved from the model's capability catalog entry.
1129+ /// </summary>
1130+ [ JsonPropertyName ( "maxPromptTokens" ) ]
1131+ public int ? MaxPromptTokens { get ; set ; }
1132+
1133+ /// <summary>
1134+ /// Overrides the maximum context window size in tokens.
1135+ /// When set, takes precedence over the default limit resolved from the model's capability catalog entry.
1136+ /// </summary>
1137+ [ JsonPropertyName ( "maxContextWindowTokens" ) ]
1138+ public int ? MaxContextWindowTokens { get ; set ; }
1139+
1140+ /// <summary>
1141+ /// Specifies the model ID used to look up default token limits from the capability catalog.
1142+ /// When unset, the session's configured model ID (see <see cref="SessionConfig.Model"/>) is used.
1143+ /// This is useful for fine-tuned models that share the same limits as a base model.
1144+ /// </summary>
1145+ [ JsonPropertyName ( "modelLimitsId" ) ]
1146+ public string ? ModelLimitsId { get ; set ; }
11181147}
11191148
11201149/// <summary>
Original file line number Diff line number Diff line change @@ -601,6 +601,19 @@ type ProviderConfig struct {
601601 BearerToken string `json:"bearerToken,omitempty"`
602602 // Azure contains Azure-specific options
603603 Azure * AzureProviderOptions `json:"azure,omitempty"`
604+ // MaxOutputTokens overrides the maximum number of output tokens the model can generate.
605+ // When set, takes precedence over the default limit from the model's capability catalog entry.
606+ MaxOutputTokens int `json:"maxOutputTokens,omitempty"`
607+ // MaxPromptTokens overrides the maximum number of prompt/input tokens.
608+ // When set, takes precedence over the default limit from the model's capability catalog entry.
609+ MaxPromptTokens int `json:"maxPromptTokens,omitempty"`
610+ // MaxContextWindowTokens overrides the maximum context window size in tokens.
611+ // When set, takes precedence over the default limit from the model's capability catalog entry.
612+ MaxContextWindowTokens int `json:"maxContextWindowTokens,omitempty"`
613+ // ModelLimitsId specifies the model ID used to look up default token limits from the capability catalog.
614+ // When unset, the session's configured model ID is used.
615+ // Useful for fine-tuned models that share the same limits as a base model.
616+ ModelLimitsId string `json:"modelLimitsId,omitempty"`
604617}
605618
606619// AzureProviderOptions contains Azure-specific provider configuration
Original file line number Diff line number Diff line change @@ -1260,6 +1260,31 @@ export interface ProviderConfig {
12601260 */
12611261 apiVersion ?: string ;
12621262 } ;
1263+
1264+ /**
1265+ * Overrides the maximum number of output tokens the model can generate.
1266+ * When set, takes precedence over the default limit resolved from the model's capability catalog entry.
1267+ */
1268+ maxOutputTokens ?: number ;
1269+
1270+ /**
1271+ * Overrides the maximum number of prompt/input tokens.
1272+ * When set, takes precedence over the default limit resolved from the model's capability catalog entry.
1273+ */
1274+ maxPromptTokens ?: number ;
1275+
1276+ /**
1277+ * Overrides the maximum context window size in tokens.
1278+ * When set, takes precedence over the default limit resolved from the model's capability catalog entry.
1279+ */
1280+ maxContextWindowTokens ?: number ;
1281+
1282+ /**
1283+ * Specifies the model ID used to look up default token limits from the capability catalog.
1284+ * When unset, the session's configured model ID is used.
1285+ * This is useful for fine-tuned models that share the same limits as a base model.
1286+ */
1287+ modelLimitsId ?: string ;
12631288}
12641289
12651290/**
Original file line number Diff line number Diff line change @@ -1962,6 +1962,16 @@ def _convert_provider_to_wire_format(
19621962 wire_azure ["apiVersion" ] = azure ["api_version" ]
19631963 if wire_azure :
19641964 wire_provider ["azure" ] = wire_azure
1965+ if "max_output_tokens" in provider :
1966+ wire_provider ["maxOutputTokens" ] = provider ["max_output_tokens" ]
1967+ if "max_prompt_tokens" in provider :
1968+ wire_provider ["maxPromptTokens" ] = provider ["max_prompt_tokens" ]
1969+ if "max_context_window_tokens" in provider :
1970+ wire_provider ["maxContextWindowTokens" ] = provider [
1971+ "max_context_window_tokens"
1972+ ]
1973+ if "model_limits_id" in provider :
1974+ wire_provider ["modelLimitsId" ] = provider ["model_limits_id" ]
19651975 return wire_provider
19661976
19671977 def _convert_custom_agent_to_wire_format (
Original file line number Diff line number Diff line change @@ -507,6 +507,19 @@ class ProviderConfig(TypedDict, total=False):
507507 # Takes precedence over api_key when both are set.
508508 bearer_token : str
509509 azure : AzureProviderOptions # Azure-specific options
510+ # Overrides the maximum number of output tokens the model can generate.
511+ # Takes precedence over the default limit from the model's capability catalog entry.
512+ max_output_tokens : int
513+ # Overrides the maximum number of prompt/input tokens.
514+ # Takes precedence over the default limit from the model's capability catalog entry.
515+ max_prompt_tokens : int
516+ # Overrides the maximum context window size in tokens.
517+ # Takes precedence over the default limit from the model's capability catalog entry.
518+ max_context_window_tokens : int
519+ # Model ID used to look up default token limits from the capability catalog.
520+ # When unset, the session's configured model ID is used.
521+ # Useful for fine-tuned models that share the same limits as a base model.
522+ model_limits_id : str
510523
511524
512525class SessionConfig (TypedDict , total = False ):
You can’t perform that action at this time.
0 commit comments