[models-bot]

Auto-generated by poc-agent for provider mistral-ai.

---

Note

Low Risk
Provider metadata-only change with no runtime logic or auth/data handling impact.

Overview
Updates the zai-glm-5-3 provider catalog entry so routing and clients can treat it as a reasoning model.

Adds thinking: true and a reasoning_effort param (default low, allowed low / high / max), aligned with other Mistral GLM/thinking model YAMLs.

^{Reviewed by Cursor Bugbot for commit d04689b. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

SDK live feature checks

Result: ✅ passed (13 scenarios)

Informational — does not block merge.

Model	Pass	Fail
✅ mistral-ai/zai-glm-5-3	13	0

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d04689b. Configure here.}

[cursor]

Incorrect reasoning effort default

Medium Severity

reasoning_effort uses defaultValue low, which Mistral does not document for zai-glm-5-3. Official GLM-5.3 specs default omitted effort to max, and this host publishes no separate default. Catalog clients will treat low as the API default.

 

^{Triggered by project rule: Provider default limits}

^{Reviewed by Cursor Bugbot for commit d04689b. Configure here.}