Handling non-OpenAI APIs

async-openai/src/types/responses/response.rs

@@ -1440,6 +1440,7 @@ pub struct ResponseLogProb {
 #[derive(Debug, Serialize, Deserialize, Clone, PartialEq)]
 pub struct OutputTextContent {
     /// The annotations of the text output.
+    #[serde(default)]
     pub annotations: Vec<Annotation>,
     pub logprobs: Option<Vec<LogProb>>,
     /// The text output from the model.
@@ -2349,6 +2350,35 @@ pub struct MCPApprovalRequest {
     pub server_label: String,
 }
 
+#[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Default)]
+pub struct InputTokenDetails {
+    /// The number of tokens that were retrieved from the cache.
+    /// [More on prompt caching](https://platform.openai.com/docs/guides/prompt-caching).
+    pub cached_tokens: u32,
+}
+
+#[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Default)]
+pub struct OutputTokenDetails {
+    /// The number of reasoning tokens.
+    pub reasoning_tokens: u32,
+}
+
+/// Usage statistics for a response.
+#[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Default)]
+#[serde(default)]
+pub struct ResponseUsage {
+    /// The number of input tokens.
+    pub input_tokens: u32,
+    /// A detailed breakdown of the input tokens.
+    pub input_tokens_details: InputTokenDetails,
+    /// The number of output tokens.
+    pub output_tokens: u32,
+    /// A detailed breakdown of the output tokens.
+    pub output_tokens_details: OutputTokenDetails,
+    /// The total number of tokens used.
+    pub total_tokens: u32,
+}
+
 #[derive(Debug, Serialize, Deserialize, Clone, PartialEq)]
 #[serde(untagged)]
 pub enum Instructions {