5912 open ai request fix

Sources/Conduit/Providers/OpenAI/OpenAIProvider+Helpers.swift

@@ -125,22 +125,33 @@ extension OpenAIProvider {
 
         // Add generation config
         if let maxTokens = config.maxTokens {
-            body["max_tokens"] = maxTokens
+            // o-series reasoning models and gpt-5+ reject "max_tokens"; use the newer key.
+            let tokenKey = requiresMaxCompletionTokens(model: model) ? "max_completion_tokens" : "max_tokens"
+            body[tokenKey] = maxTokens
         }
 
-        body["temperature"] = config.temperature.openAIJSONNumber
-        body["top_p"] = config.topP.openAIJSONNumber
+        // o-series and gpt-5+ only accept default values for these sampling params; omit them entirely.
+        if requiresMaxCompletionTokens(model: model) {
+            let defaults = GenerateConfig.default
+            if config.temperature != defaults.temperature || config.topP != defaults.topP
+                || config.frequencyPenalty != defaults.frequencyPenalty || config.presencePenalty != defaults.presencePenalty {
+                logger.debug("Model '\(model.rawValue)' does not support custom sampling parameters; temperature, top_p, and penalties will be ignored.")
+            }
+        } else {
+            body["temperature"] = config.temperature.openAIJSONNumber
+            body["top_p"] = config.topP.openAIJSONNumber
 
-        if let topK = config.topK {
-            body["top_k"] = topK
-        }
+            if let topK = config.topK {
+                body["top_k"] = topK
+            }
 
-        if config.frequencyPenalty != 0 {
-            body["frequency_penalty"] = config.frequencyPenalty.openAIJSONNumber
-        }
+            if config.frequencyPenalty != 0 {
+                body["frequency_penalty"] = config.frequencyPenalty.openAIJSONNumber
+            }
 
-        if config.presencePenalty != 0 {
-            body["presence_penalty"] = config.presencePenalty.openAIJSONNumber
+            if config.presencePenalty != 0 {
+                body["presence_penalty"] = config.presencePenalty.openAIJSONNumber
+            }
         }
 
         if !config.stopSequences.isEmpty {
@@ -256,8 +267,16 @@ extension OpenAIProvider {
         if let maxTokens = config.maxTokens {
             body["max_output_tokens"] = maxTokens
         }
-        body["temperature"] = config.temperature.openAIJSONNumber
-        body["top_p"] = config.topP.openAIJSONNumber
+
+        if requiresMaxCompletionTokens(model: model) {
+            let defaults = GenerateConfig.default
+            if config.temperature != defaults.temperature || config.topP != defaults.topP {
+                logger.debug("Model '\(model.rawValue)' does not support custom sampling parameters; temperature and top_p will be ignored.")
+            }
+        } else {
+            body["temperature"] = config.temperature.openAIJSONNumber
+            body["top_p"] = config.topP.openAIJSONNumber
+        }
 
         if !config.stopSequences.isEmpty {
             body["stop"] = config.stopSequences
@@ -576,6 +595,15 @@ extension OpenAIProvider {
         }
     }
 
+    /// Returns true for models (o-series, gpt-5+) that require `max_completion_tokens`
+    /// and reject non-default sampling params (temperature, top_p, penalties).
+    private nonisolated func requiresMaxCompletionTokens(model: ModelIdentifier) -> Bool {
+        let raw = model.rawValue
+        // Strip optional "provider/" prefix used by OpenRouter (e.g. "openai/gpt-5-mini" → "gpt-5-mini").
+        let id = raw.firstIndex(of: "/").map { String(raw[raw.index(after: $0)...]) } ?? raw
+        return id.hasPrefix("o1") || id.hasPrefix("o3") || id.hasPrefix("o4") || id.hasPrefix("gpt-5")
+    }
+
     /// Serializes reasoning configuration.
     private nonisolated func serializeReasoningConfig(_ config: ReasoningConfig) -> [String: Any] {
         var result: [String: Any] = [:]