Hello, I experienced something odd while using the Batch API for embeddings. What happened: I ran multiple batches of 2000 embedding requests. When I monitor the API platform console, I see: 2000 completed, 0 failed of 2000 total requests …but the generated error file actually contained sever…

I got similar behavior using chat completions today, and I’m still trying to debug it. This is my batch information from the API Console: Created at: Nov 1, 2025, 12:17 AM Endpoint: /v1/chat/completions Completion window: 24h Completion time: 18 minutes Request counts: 200 completed, 0 failed of 2…

I am seeing it too. Running batch queries on gpt-4o-mini.

I don’t think it is similar. The issue is that batch output status is Completed, and I see that 2000 completed, 0 failed of 2000 total requests. However, I see some responses are dropped in the output file and those dropped request can be found in the error file. The message I see (2000 completed, 0…

Batch API shows “2000 completed, 0 failed” — but some requests return status_code 0

API Bugs

OpenAI_Support November 15, 2025, 12:40am 4

Hello!

This issue might be similar to the one tracked here: Batch stuck in “finalizing” status for over 24 hours

Let me know if not though - happy to debug further.

Topic		Replies	Views
Batch API shows incorrect number of failed responses Bugs api	1	105	November 7, 2025
Batch API consistently expires without completion despite valid JSONL - Need investigation Bugs batch-api	0	147	November 18, 2025
Some batches creation FAILED even though they were within the batch queue limit API embeddings , batch-api	0	422	December 23, 2024
Inconsistent Number of Entries in JSONL Files from OpenAI Batch API API batch-api	1	193	June 5, 2025
Batch API requests keep falling "failed" after "finalizing" and costing my credits Bugs	1	844	August 5, 2024

Batch API shows “2000 completed, 0 failed” — but some requests return status_code 0

Related topics