Skip to content

feat: parse and return token usage in Chat Completions stream #7000

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
448523760 wants to merge 3 commits into
base: main
Choose a base branch
from
Open

feat: parse and return token usage in Chat Completions stream #7000

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
c5be851
feat: parse and return token usage in Chat Completions stream
448523760 Nov 26, 2025
950f75d
fixup
448523760 Nov 26, 2025
078af36
fixup code format
448523760 Nov 26, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
58 changes: 54 additions & 4 deletions codex-rs/codex-api/src/sse/chat.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@ use codex_client::StreamResponse;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ReasoningItemContent;
use codex_protocol::models::ResponseItem;
use codex_protocol::protocol::TokenUsage;
use eventsource_stream::Eventsource;
use futures::Stream;
use futures::StreamExt;
Expand Down Expand Up @@ -50,6 +51,8 @@ pub async fn process_chat_sse<S>(
let mut assistant_item: Option<ResponseItem> = None;
let mut reasoning_item: Option<ResponseItem> = None;
let mut completed_sent = false;
let mut token_usage: Option<TokenUsage> = None;
let mut response_id = String::new();

loop {
let start = Instant::now();
Expand Down Expand Up @@ -78,8 +81,8 @@ pub async fn process_chat_sse<S>(
if !completed_sent {
let _ = tx_event
.send(Ok(ResponseEvent::Completed {
response_id: String::new(),
token_usage: None,
response_id: response_id.clone(),
token_usage: token_usage.clone(),
}))
.await;
}
Expand Down Expand Up @@ -110,6 +113,16 @@ pub async fn process_chat_sse<S>(
}
};

if response_id.is_empty()
&& let Some(id) = value.get("id").and_then(serde_json::Value::as_str)
{
response_id = id.to_string();
}

if let Some(usage) = value.get("usage") {
token_usage = parse_openai_usage(usage);
}

let Some(choices) = value.get("choices").and_then(|c| c.as_array()) else {
continue;
};
Expand Down Expand Up @@ -201,8 +214,8 @@ pub async fn process_chat_sse<S>(
if !completed_sent {
let _ = tx_event
.send(Ok(ResponseEvent::Completed {
response_id: String::new(),
token_usage: None,
response_id: response_id.clone(),
token_usage: token_usage.clone(),
}))
.await;
completed_sent = true;
Expand Down Expand Up @@ -297,6 +310,43 @@ async fn append_reasoning_text(
}
}

fn parse_openai_usage(usage: &serde_json::Value) -> Option<TokenUsage> {
let prompt_tokens = usage
.get("prompt_tokens")
.and_then(serde_json::Value::as_i64)
.unwrap_or(0);

let completion_tokens = usage
.get("completion_tokens")
.and_then(serde_json::Value::as_i64)
.unwrap_or(0);

let total_tokens = usage
.get("total_tokens")
.and_then(serde_json::Value::as_i64)
.unwrap_or(0);

let cached_input_tokens = usage
.get("prompt_tokens_details")
.and_then(|d| d.get("cached_tokens"))
.and_then(serde_json::Value::as_i64)
.unwrap_or(0);

let reasoning_output_tokens = usage
.get("completion_tokens_details")
.and_then(|d| d.get("reasoning_tokens"))
.and_then(serde_json::Value::as_i64)
.unwrap_or(0);

Some(TokenUsage {
input_tokens: prompt_tokens,
cached_input_tokens,
output_tokens: completion_tokens,
reasoning_output_tokens,
total_tokens,
})
}

#[cfg(test)]
mod tests {
use super::*;
Expand Down
Loading