Hi @siddhant , I am joining the long chorus of those who would be very grateful for your help in getting past the batch issue. My affected batches are:
batch_6a976599106881908b64710f2cddb261 |
failed |
|---|---|
batch_6a9765cd091c8190b230be6dd6c9aaed |
failed |
batch_6a9765f4e2b88190852d72bb5e5eb8c4 |
failed |
Thank you so much in advance for any help
It just works now, thank you and the team.
Hey, all. Thanks to everyone reporting in this thread.
If you have time, I recommend checking out this thread…
… and sticking around.
We’ve got a great community here and are always looking for new people!
@vanessa2,@filip.kostecki- done
@siddhant Would appreciate you patching the following batches. Thanks!
| batch_6a97d7e04d1c81908e6f1d369838f706 | 2026/09/02 08:01:36 UTC |
| batch_6a97e2cb2c5481908573ad820e1b1a19 | 2026/09/02 08:48:11 UTC |
| batch_6a97e2c259ac819081c636c8bd44ab03 | 2026/09/02 08:48:02 UTC |
| batch_6a97e1e0424081909052f95b2bf5947d | 2026/09/02 08:44:16 UTC |
| batch_6a97e0c8d6608190a61f87558a61eb26 | 2026/09/02 08:39:36 UTC |
| batch_6a97dfd4c9cc8190800eb95027d3b04a | 2026/09/02 08:35:32 UTC |
| batch_6a97de9e44b48190a03f523f83e9bea9 | 2026/09/02 08:30:22 UTC |
| batch_6a97dd45e6448190a0d8d1c41011ac25 | 2026/09/02 08:24:37 UTC |
| batch_6a97dc30bea881909d584708efb30796 | 2026/09/02 08:20:00 UTC |
| batch_6a97daedc24081909c68c414d48b1fbd | 2026/09/02 08:14:37 UTC |
| batch_6a97d9ebfc508190b0ef438360ea9d6d | 2026/09/02 08:10:19 UTC |
| batch_6a97d8f8e93081908be35681d4f1e4f9 | 2026/09/02 08:06:16 UTC |
@Samme,@vishalvarwani8-done
Also affected, since 27 Aug 2026. 100% of our batches fail, no successes since.
Per the note at the top of this thread, no org ID below, just failed batch IDs and submission timestamps with timezone.
Failed batches. All times Europe/Zagreb, CEST, UTC+2
| Batch ID | Submitted |
|---|---|
batch_6a97eb06b0cc8190b5527e72145a245e |
2026-09-02 11:23:18 CEST / 09:23:18 UTC, failed 11:23:45 CEST, 27s later |
batch_6a983f94c8c48190b89622e3f0b62439 |
2026-09-02 17:24 CEST |
batch_6a97b73a8e488190b63d1ff764613efa |
2026-09-02 07:42 CEST |
batch_6a97b6ae443881909da344c535b7d926 |
2026-09-02 07:39 CEST |
batch_6a97b61341dc8190afda41febfdc1b7a |
2026-09-02 07:37 CEST |
batch_6a97b548d65c8190adcd6948d8965b4e |
2026-09-02 07:34 CEST |
Same error as the OP, with request_counts 0 completed / 0 failed / 0 total, so the input file is never read.
What we verified on our side
/v1/responses, completion_window: 24hpurpose=batch, not expired. I downloaded it back and re-parsed it: valid single-line JSONL with correct custom_id, method, url and body./v1/responses call with the same API key, so the key, the model and the request body are all fine.new OpenAI({ apiKey }) with no project, no organization, no OpenAI-Project header and no custom baseURL. The same client instance calls files.create({ purpose: "batch" }) and then batches.create({ input_file_id, endpoint: "/v1/responses" }) as sequential awaits.openai-node 6.3.0, unchanged. Our last successful batch was 18 Aug 2026 and nothing changed on our side in between.So the usual “the file was uploaded under a different project or key context” explanation does not apply here: same client, same key, same project, and the batch still cannot resolve its own input file.
One thing worth flagging: the AI-assisted support chat told us today that it sees no active incident on the status page matching this, and that it could not confirm account-level impact or give an ETA. That contradicts the note at the top of this thread about temporary fixes being applied to affected organizations. An incident ID we could quote to support would help.
Could our org be added to the temporary fix list? Happy to provide anything further that does not include the org ID.
Hi @siddhant , I’m still getting the same issue. Our org has the same Batch API issue. We discovered the issue just now. Could you please apply the fix to our organization?
Batch id:
Thanks for your help.
Hi @siddhant could you fix our org:
batch_6a97fc3ad55481909fc72760f3456239 and timestamp
2 Sept 2026, 11:36 (UK time)
Hi @siddhant, could you apply the fix to our organization? Failed batches, times in CEST (UTC+2):
batch_6a915fa8aeb48190adbe98c7a4aa9ffd 2026-08-28 12:15:04 CEST / 10:15:04 UTC
batch_6a97fe2e63dc81908dc2639a2a5fa02c 2026-09-02 12:45:02 CEST / 10:45:02 UTC
batch_6a9801b318288190a7b4a0cb776a8e6c 2026-09-02 13:00:03 CEST / 11:00:03 UTC
batch_6a980536b37881908201261132f6b160 2026-09-02 13:15:02 CEST / 11:15:02 UTC
100% failure since 2026-08-28, no successes. Thanks!
Hi @siddhant It seems we are also facing this issue. Could you take a look? Thank you in advance 
batch_6a94c409ee288190a8a506099ed4b43e | Aug 31, 2026, 3:00 AM
batch_6a9767835d008190b13beb21da3a0ed9 | Sep 2, 2026, 3:02 AM
hi @siddhant , can you please fix org associated with the batches provided in my message above? Thanks a lot!
We are also affected, including on /v1/chat/completions. Could you please investigate this failed Batch and check whether the temporary fix mentioned at the top of this topic applies to our organization?
batch_6a98517a918881908eb224c58c6e6503created_at: 2026-09-02 16:40:26 UTC (2026-09-03 00:40:26 UTC+08:00)./v1/chat/completions; completion window: 24h.gpt-5.6-sol. We also reproduced the same file-access failure in a separate minimal test using gpt-5.5.failed; error code: invalid_request; parameter: file_id.Cannot find file file-..., or organization org-... does not have access to it.completed=0, failed=0, total=0. The failure occurs during validation, before any model request is processed.Our latest reproduction used one official Python openai.OpenAI client (SDK 2.3.0) and the same project-scoped API key for upload, file reads, and Batch creation. No organization or project override headers were sent. We also reproduced the issue separately with direct REST calls and explicit organization/project scope.
The newly uploaded input file had purpose=batch and status=processed. Both metadata retrieval and content download succeeded with HTTP 200 before Batch creation and again after the Batch failed. The downloaded bytes before and after failure matched. The Batch used the exact file ID returned by this upload; it was not an old or expired file.
The input contained just one nonclinical text request asking the model to Reply with exactly OK., with reasoning_effort=none and max_completion_tokens=128. There were no images, attachments, or embedded file references.
The Files API can therefore read the input file using the same credential while Batch validation reports that it cannot access that same file. We cannot determine the internal cause from client-side evidence alone. Please let us know whether you can apply the relevant mitigation or need additional private diagnostics. Organization/project IDs and request IDs can be provided through a private support channel.
Hi @siddhant
Same failure here, since 2026-09-02 10:35 UTC. Last successful batches for this org were on 2026-08-02.
Every batch we create goes validating → failed with: “Cannot find file file-…, or organization org-… does not have access to it.”
The input files upload fine (HTTP 200, purpose=batch, status=processed), and the same API key can list, retrieve and download them. Only the batch validator can’t see them.
What we tried, all with the identical failure: waiting for status=processed plus a delay before batches.create; re-creating the batch from the same file id repeatedly (over 6 hours, 15-min intervals); a 1-request test file; an explicit OpenAI-Organization header; uploading via the Uploads API (/v1/uploads → parts → complete) instead of /v1/files.
Recent failed batch IDs (endpoint /v1/chat/completions, model gpt-5.4-mini):
batch_6a9820d2c6a08190a631da746e724dbf 2026-09-02 13:13 UTC
batch_6a98208a1e7c8190979802b07e14e0e4 2026-09-02 13:12 UTC (1-request test file)
batch_6a98436342388190b73210442b8dca5e 2026-09-02 15:41 UTC
batch_6a985906078c819083b170189af76123 2026-09-02 17:12 UTC
Could you apply the per-org fix to us as well? Happy to provide more batch IDs or run a test batch on request.
Akash
Hi @siddhant , could you apply the fix to our organization? 100% failure since Aug 28, all failing at validation with “Cannot find file X, or organization does not have access to it” despite files.retrieve showing status: processed.
Failed batches (UTC):
Thanks!
We are having the same issue @siddhant Our latest test batch which failed is:
batch_6a984e9755b88190b57466aaabbd6bcf - UTC 4:28:07 PM, 2026-09-02
Could the fix please be applied to our organization? Thank you
batch_6a984393250c819080e4e438a45d75bc
Thanks
Hi all 
We’ve now deployed the permanent fix for the Batch API input-file access issue and removed the temporary workarounds.
If we previously applied a workaround for your organisation and you start seeing the exact same error again, please reply in this thread with the batch ID, the time the error occurred, and your timezone.
Sorry again for the disruption, and thank you for your patience while we worked through this
Sep 2, 2026, 11:56 AM, MDT
batch_6a9845d43270819082e91e3c04555252
batch_6a9845d44d8881909413b35f06672189
batch_6a9845d4aae48190870ca82dbfbd238a
batch_6a9846d576448190825936c32e0bb214
batch_6a9846d5d0cc8190867355c2872e8be6
batch_6a9846d5d2788190a589f9b4c8ef58a5
batch_6a984767feb0819099cc6efb538eb3db
batch_6a98483a9268819097fc016acaea3cc1
batch_6a985205bbec81908d62d6914c8455dc
batch_6a9853ed97a48190b8830878a36c0939
batch_6a986369fa548190af3dc4121e89f579
Hey, all. Thanks to everyone reporting in this thread.
If you have time, I recommend checking out this thread…
… and sticking around.
We’ve got a great community here and are always looking for new people!