anomalyco
Using `GPT 5.1 Codex` via OpenAI fails on the 2nd request
Description
┃ # Summarizing provided text to summarize 14,515/4% ($0.02) ┃
┃
┃ Hi
┃ varungawande (12:46:42 PM)
┃
Hello! How can I help today?
Build gpt-5.1-codex
┃
┃ Hello
┃ varungawande (12:46:44 PM)
┃
┃
┃ Item with id 'rs_04819b77d3ae6f3101691c1d5bb4508190a6bff984579fc868' not found. Items are not persisted when `store` is set to false. Try again with `store` set to true, or remove this item from your
┃ input.
┃
Item with id 'rs_04819b77d3ae6f3101691c1d5bb4508190a6bff984579fc868' not found. Items are not persisted when `store` is set to false. Try again with `store` set to true, or remove this item from your input.
When I use this model via OpenRouter it works properly.
OpenCode version
1.0.72
Steps to reproduce
Using GPT-5.1 Codex via OpenAI
- Send "Hi" -> Works
- Send "Hello" next -> Fails
Note: My config has nothing related to this model:
{
"$schema": "https://opencode.ai/config.json",
"share": "disabled",
"permission": {
"bash": {
"git commit": "ask",
"git status*": "allow",
"git log*": "allow",
"git diff*": "allow",
"git push": "deny",
"git reset": "ask",
"git rebase": "ask",
"git merge": "ask",
"git cherry-pick": "ask",
"sudo": "ask",
"rm": "ask",
"mv": "ask",
"chmod": "ask",
"chown": "ask"
}
}
Screenshot and/or share link
Operating System
Darwin Kernel Version 25.0.0: Wed Sep 17 21:41:39 PDT 2025; root:xnu-12377.1.9~141/RELEASE_ARM64_T8103 arm64
Terminal
Ghostty 1.2.3
Like this?
Didn't work:
"provider": {
"openai": {
"models": {
"gpt-5.1-codex": {
"name": "GPT-5.1 Codex (high)",
"options": {
"reasoningEffort": "high",
"store": false
}
}
}
}
}
}
Or if you meant this, that didn't work either:
"gpt-5.1-codex": {
"name": "GPT-5.1 Codex (high)",
"options": {
"reasoningEffort": "high"
},
"store": false
},
@V4G4X try this one: "gpt-5.1-codex": { "name": "GPT-5.1 Codex (high)", "options": { "include": ["reasoning.encrypted_content"], "store": false, }, },
I'm also running into this: "Item with id rs_* not found"
I've tried the store false change and seeing it show up randomly still.
I'm trying this with azure/gpt-5.1-codex-max. Since the fix didn't work for me and it's impacted almost every session at some point, figured I'd make sure it was brought up again.
@sheldonhull can u share a session by chance?
opencode export > session.json
Hopefully this is enough, since otherwise it's a lot of stuff to obfuscate ...
log details
{
"info": {
"id": "msg_b2e9f755d0017W7BiVIhsSw62C",
"sessionID": "ses_4d1693c9bffeNyPRnoPsGaMcks",
"role": "assistant",
"time": {
"created": 1766013760861,
"completed": 1766013765998
},
"parentID": "msg_b2e96c3670015vDk5DIyVf3US9",
"modelID": "gpt-5.1-codex-max",
"providerID": "azure",
"mode": "build",
"path": {
"cwd": "/Users/{PATH}daggerverse",
"root": "/Users/{PATH}daggerverse"
},
"cost": 0.01465175,
"tokens": {
"input": 219,
"output": 193,
"reasoning": 128,
"cache": {
"read": 89344,
"write": 0
}
},
"finish": "tool-calls"
},
"parts": [
{
"id": "prt_b2e9f78f4001eZ6xqb510jYtSM",
"sessionID": "ses_4d1693c9bffeNyPRnoPsGaMcks",
"messageID": "msg_b2e9f755d0017W7BiVIhsSw62C",
"type": "step-start",
"snapshot": "ccde326bfdf64c680e2095f16e7447131bcefe1c"
},
{
"id": "prt_b2e9f7eaf001FpKAFHxuVWOknu",
"sessionID": "ses_4d1693c9bffeNyPRnoPsGaMcks",
"messageID": "msg_b2e9f755d0017W7BiVIhsSw62C",
"type": "reasoning",
"text": "",
"metadata": {
"openai": {
"itemId": "rs_0f6b4bc48d7a18620069433b433f1081968699460f3a33bde6",
"reasoningEncryptedContent": null
}
},
"time": {
"start": 1766013763247,
"end": 1766013764899
}
},
{
"id": "prt_b2e9f8524001wwVUiK6u7zkAPD",
"sessionID": "ses_4d1693c9bffeNyPRnoPsGaMcks",
"messageID": "msg_b2e9f755d0017W7BiVIhsSw62C",
"type": "tool",
"callID": "call_aZPOp47XEn1LhZRpgDeaQvda",
"tool": "bash",
"state": {
"status": "completed",
"input": {
"command": "cd /Users/{PATH}daggerverse/pr-reviewer && gofmt -w main.go suggestion.go azdo.go",
"description": "Formats updated Go files"
},
"output": "",
"title": "Formats updated Go files",
"metadata": {
"output": "",
"exit": 0,
"description": "Formats updated Go files"
},
"time": {
"start": 1766013765720,
"end": 1766013765933
}
},
"metadata": {
"openai": {
"itemId": "fc_0f6b4bc48d7a18620069433b44e0d8819690142f18bda9c485"
}
}
},
{
"id": "prt_b2e9f89300019OV7BALuD2yoEK",
"sessionID": "ses_4d1693c9bffeNyPRnoPsGaMcks",
"messageID": "msg_b2e9f755d0017W7BiVIhsSw62C",
"type": "step-finish",
"reason": "tool-calls",
"snapshot": "ccde326bfdf64c680e2095f16e7447131bcefe1c",
"cost": 0.01465175,
"tokens": {
"input": 219,
"output": 193,
"reasoning": 128,
"cache": {
"read": 89344,
"write": 0
}
}
}
]
},
{
"info": {
"id": "msg_b2e9f8991001VU6D1XgdY8xVKs",
"sessionID": "ses_4d1693c9bffeNyPRnoPsGaMcks",
"role": "assistant",
"time": {
"created": 1766013766033,
"completed": 1766013766804
},
"error": {
"name": "APIError",
"data": {
"message": "Item with id 'rs_0f6b4bc48d7a18620069433b433f1081968699460f3a33bde6' not found.",
"statusCode": 400,
"isRetryable": false,
"responseHeaders": {
"apim-request-id": "{GUID}",
"content-length": "194",
"content-type": "application/json",
"date": "Wed, 17 Dec 2025 23:22:46 GMT",
"strict-transport-security": "max-age=31536000; includeSubDomains; preload",
"x-content-type-options": "nosniff",
"x-ms-region": "East US 2",
"x-request-id": "{GUID}"
},
"responseBody": "{\n \"error\": {\n \"message\": \"Item with id 'rs_0f6b4bc48d7a18620069433b433f1081968699460f3a33bde6' not found.\",\n \"type\": \"invalid_request_error\",\n \"param\": \"input\",\n \"code\": null\n }\n}"
}
},
"parentID": "msg_b2e96c3670015vDk5DIyVf3US9",
"modelID": "gpt-5.1-codex-max",
"providerID": "azure",
"mode": "build",
"path": {
"cwd": "/Users/{PATH}daggerverse",
"root": "/Users/{PATH}daggerverse"
},
"cost": 0,
"tokens": {
"input": 0,
"output": 0,
"reasoning": 0,
"cache": {
"read": 0,
"write": 0
}
}
},
"parts": []
},
For me this has been seriously impactful to using Opencode so I rely more on other tools right now. Nearly every conversation has this experience with gpt-5 models.
Thankfully I've been able to just use export and start in a new session easily enough but it essentially ends up blocking a current session requiring me to start a new one.
I get it with gpt 5.1 mini, max, normal, codex.
Just wanted to make sure you knew it's not just an occasional annoyance but makes it a bit rough.
I think I'd need an entire session export to actually debug this, I never run into it msyelf,
Is this happening if you abort mid session?