Azure Foundry tool sessions still fail on 0.118.0 without local payload patching

[Alfredo-Sandoval]

What version of Codex CLI is running?

codex-cli 0.118.0

What subscription do you have?

Azure Foundry / Azure OpenAI only

Which model were you using?

gpt-5.4 via Azure Foundry

What platform is your computer?

macOS arm64

What terminal emulator and version are you using (if applicable)?

zsh on macOS

What issue are you seeing?

Azure-backed Codex CLI tool sessions are still broken on 0.118.0.

I have a local shell alias named codex-azure, but that alias is only:

codex --profile azure_foundry

So this is not an alias bug. The underlying Azure profile itself is broken for tool-driven sessions.

With direct Azure Foundry configuration, Codex can start a session and even run the first tool call, but the follow-up /responses request fails with Azure:

{
  "error": {
    "code": "invalid_payload",
    "message": "The provided data does not match the expected schema",
    "param": "/",
    "type": "invalid_request_error",
    "details": [],
    "additionalInfo": {
      "request_id": "22d9b916fade3079190d29139a72da64"
    }
  }
}

In practice, Azure/Codex is not usable for normal interactive repo work unless I put a local patching proxy in front of Azure and mutate the payloads before they go upstream.

That proxy is only a workaround. Without patching, Azure tool use is still broken.

This looks related to #15584 and #14695, but it is still reproducible on 0.118.0.

What steps can reproduce the bug?

1. Configure Codex CLI to use Azure Foundry / Azure OpenAI with wire_api = "responses".
2. Run Codex with that Azure profile.
3. Ask for anything that triggers a shell tool call.

Minimal repro:

codex --profile azure_foundry exec "Run git status --short --branch and summarize the repo state in two sentences."

Equivalent local alias repro:

codex-azure exec "Run git status --short --branch and summarize the repo state in two sentences."

Observed behavior:

• the shell tool itself runs
• then the next Azure /responses request fails with invalid_payload
• interactive tool-driven work cannot continue

I also reproduced it in interactive mode by starting Codex and asking it to inspect a repo. As soon as the first tool-backed turn completes, Azure returns schema/payload errors.

What is the expected behavior?

Azure Foundry should accept the tool round-trip and allow Codex to continue the session normally, without requiring a user-side proxy that patches or strips malformed payload items.

Additional information

Relevant local setup details:

• codex-azure is just an alias/wrapper for codex --profile azure_foundry
• the bug reproduces with the underlying profile directly, so the alias is not the cause
• direct Azure Foundry endpoint: broken after tool call
• local payload-patching proxy: can partially mask the issue, which strongly suggests Codex is generating Azure-incompatible /responses continuation payloads

This means Azure support is effectively not working out of the box right now for normal tool-using Codex sessions unless users patch around it locally.

[github-actions]

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

• Tool Calls Broken when Using Azure OpenAI models with Codex #15584

Powered by Codex Action

[etraut-openai]

Thanks for the bug report. Do you have reason to believe that this is a Codex client issue? If so, was it working with a previous version of the client? For example, if you downgrade to 0.116.0, does it work?

Based on an initial read of your report, I suspect this might be an Azure service issue. If you haven't already done so, I recommend reporting this to the folks at Azure. If this is an issue on their end, they will not know about it unless you report it.

[ajboloor]

Seeing the same error on my end.

---

What version of Codex CLI is running?

codex-cli 0.118.0

What subscription do you have?

Azure Foundry / Azure OpenAI only

Which model were you using?

gpt-5.3-codex via Azure Foundry

What platform is your computer?

windows x64

What terminal emulator and version are you using (if applicable)?

bash on WSL

---

Here is one conversation:

› hello
 Hi. What do you want to work on in this repo?


› what is the weather in SFO?


• Searching the web

• Searched weather: San Francisco, CA

────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────

• For San Francisco (SFO area) today, Monday, April 6, 2026: it’s around 58°F now, with a day range near 60°F / 50°F and little to no rain expected. (a
  .msn.com (https://a.msn.com/54/en-us/ct37.780%2C-122.420))

  If you want, I can give the airport-specific SFO METAR/TAF conditions next.


› what is this repo?


■ {
  "error": {
    "code": "invalid_payload",
    "message": "The provided data does not match the expected schema",
    "param": "/",
    "type": "invalid_request_error",
    "details": [],
    "additionalInfo": {
      "request_id": "7a57e9927fca7fd75a4328c3adb45041"
    }
  }
}

> hello
■ {
  "error": {
    "code": "invalid_payload",
    "message": "The provided data does not match the expected schema",
    "param": "/",
    "type": "invalid_request_error",
    "details": [],
    "additionalInfo": {
      "request_id": "722b5ec74bbfa9fdc25388204c957be7"
    }
  }
}

Some additional info:

• The failure seems to begin during particular tool calls. It is interesting that once the failure gets triggered when a coding based tool is called, it will continue to fail even when there is no tool call.
• Same issue on codex-cli 0.116.0.
• I was following the instructions from https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/codex
• Restarting the codex session resolves the issue until a particular tool call is invoked.

[minimAluminiumalism]

Same error here. But I found azure gpt is working well in opencode. It doesn't look like an Azure issue?