Spark model rejects reasoning.summary with unsupported_parameter

[ozan-s]

What version of Codex CLI is running?

codex-cli 0.106.0

What subscription do you have?

ChatGPT Pro

Which model were you using?

gpt-5.3-codex-spark

What platform is your computer?

macOS

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

Not provided

What issue are you seeing?

Selecting the Spark model fails immediately with:

{
  "type": "error",
  "error": {
    "type": "invalid_request_error",
    "code": "unsupported_parameter",
    "message": "Unsupported parameter: 'reasoning.summary' is not supported with the 'gpt-5.3-codex-spark' model.",
    "param": "reasoning.summary"
  },
  "status": 400
}

What steps can reproduce the bug?

1. Select gpt-5.3-codex-spark in Codex CLI.
2. Submit a normal prompt.
3. Request fails with unsupported_parameter on reasoning.summary.

What is the expected behavior?

Codex should not send reasoning.summary for models that do not support it, or Spark should advertise/support the parameter consistently.

Additional information

Findings from local repo analysis suggest this may be a recent capability mismatch/regression window:

• Codex sends reasoning.summary only when model metadata indicates reasoning summaries are supported.
• On Feb 26, 2026, commit ba41e84a5 changed summary fallback behavior to rely on model-catalog defaults.
• On Feb 25, 2026, commit d45ffd583 updated 5.3 model visibility/metadata in bundled model data.

Hypothesis: Spark capability metadata/default summary handling and backend parameter support are out of sync, causing reasoning.summary to be sent for a model that rejects it.

[github-actions]

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

• /review can submit stale reasoning.effort after switching to gpt-5.1-codex-mini #12495

Powered by Codex Action

[ledmaster]

Found the same issue.

I have custom "fast review" agents based on gpt-5.3-codex-spark and since updating to v0.106.0, the subagents fail due to:

{ "type": "error", "error": { "type": "invalid_request_error", "code": "unsupported_parameter", "message": "Unsupported parameter: 'reasoning.summary' is not supported with the 'gpt-5.3- codex-spark' model.", "param": "reasoning.summary" }, "status": 400 }

[pakrym-oai]

Do you happen to have a model_reasoning_summary setting in the config?

[ozan-s]

Yes, I have this line in my config:
model_reasoning_summary = "detailed"

[ledmaster]

Do you happen to have a model_reasoning_summary setting in the config?

model_reasoning_summary_format="detailed"
model_reasoning_summary="detailed"

[pakrym-oai]

After the recent change the CLI started respecting the setting. Before that it always used auto.

[ledmaster]

After the recent change the CLI started respecting the setting. Before that it always used auto.

Is there a way to specify this config by model or is the recommended action to set it to auto?

Thanks for the quick replies

[pakrym-oai]

I recommend removing the config all together, then the most optimal model setting will be used.

I expect you won't get much more value out of detailed reasoning summaries.

[dbathily]

I reproduced this again and there seems to be a narrower split in behavior.

I removed:

model_reasoning_summary = "detailed"

from ~/.codex/config.toml.

After that:

• explicit spawn_agent(..., model="gpt-5.3-codex-spark") worked in a simple read-only smoke test
• but role-based subagents still failed when their .codex/agents/*.toml had:

model = "gpt-5.3-codex-spark"
model_reasoning_effort = "high"

Affected role configs in my repo:

• .codex/agents/explorer.toml
• .codex/agents/service_owner_investigator.toml
• .codex/agents/integration_investigator.toml

Role-based spawn still fails with:

{
  "type": "error",
  "error": {
    "type": "invalid_request_error",
    "code": "unsupported_parameter",
    "message": "Unsupported parameter: 'reasoning.summary' is not supported with the 'gpt-5.3-codex-spark' model.",
    "param": "reasoning.summary"
  },
  "status": 400
}

So removing model_reasoning_summary fixes the explicit model path for me, but not the custom-role/subagent path. It looks like role-based spawns may still be attaching reasoning.summary somewhere separately.

[Morriz]

yes, removing it still seems to persist somewhere