OpenAI API · api · fg-0142

Streamed chat completions report no token usage, so cost tracking records zero

chunk.usage is null on every streamed chunk

Verified 2026-10-15 · Confirmed by 0 builders · Trust 50/100

Cause

Streaming responses omit usage unless stream_options.include_usage is set, and then only a final extra chunk with empty choices carries the totals. An interrupted stream never sends it.

Fix

Set include_usage, read usage from the last chunk, and guard code that indexes choices[0].

Apply

const stream = await client.chat.completions.create({ model, messages, stream: true, stream_options: { include_usage: true } });
for await (const chunk of stream) {
  if (chunk.usage) recordUsage(chunk.usage);
  const delta = chunk.choices[0]?.delta?.content;
}

Affects: OpenAI Chat Completions streaming, checked October 2026

Sources

• https://github.com/openai/openai-python/blob/main/src/openai/types/chat/chat_completion_stream_options_param.py

claude mcp add --transport http footgun https://getfootgun.com/api/mcp