# Some GPT 5.2 calls failing with inference issues

- Provider: [Vapi](https://down.guide/providers/vapi)
- Confidence: Confirmed
- Impact: Partial outage
- lasted 9d 6h
- started 2026-03-10 12:46 UTC
- id: DWN-95e835a34d
- Official page: https://status.vapi.ai/incident/846049

## Updates from Vapi
### 2026-03-10 12:46 UTC: update
We're noticing a small percentage of calls using GPT 5.2 fail with an internal error during inference from OpenAI's side. We've reached out to the team and are closely monitoring the situation. In the meantime, we recommend switching to another model as we're seeing the degradation only on 5.2 currently.

### 2026-03-19 18:51 UTC: resolved
The issue has been resolved we're not seeing any further degradation

From the builder: does your app depend on Vapi? [Sazabi](https://www.sazabi.com/?utm_source=down-guide&utm_medium=markdown&utm_content=vapi) detects problems in your own logs without hand-written monitors, and you can ask it in plain language which errors mention Vapi.

---
down.guide reads providers' official status pages every minute and adds its own evidence, always labeled. Free, no account. Agents: MCP at https://down.guide/mcp, REST at https://down.guide/api/v1, feeds at https://down.guide/feed. Built by [Sazabi](https://www.sazabi.com/?utm_source=down-guide&utm_medium=markdown&utm_content=DWN-95e835a34d), AI-native observability and security: automatic issue detection, plain-language log search and live dashboards.