# KeyHunter operation harvested 2,233 LLM API keys

*Security incident · 8 Nov 2026*

SOCRadar's Threat Research Unit has published a report on the KeyHunter operation, which steals LLM API credentials from misconfigured cloud infrastructure. According to the report, published on 4 November 2026, the actor harvested 2,233 API keys. Keys from Anthropic, OpenAI, Google and DeepSeek are among them, and the actor's own tooling marked 12 as valid. The researchers also catalogued 247 exposed LLM gateways, of which 223 responded as usable.

The operation is attributed to a Chinese-speaking actor and was active as late as 13 October 2026. Its LLM key theft runs alongside an academic fraud campaign and card-validation activity.

For users of LLM APIs, the practical consequence is the same as in other key-theft cases: a leaked key can be used to run up usage on the owner's account. Rotate API keys that were stored in shared configuration, environment files or public repositories, and review the provider's usage and billing logs for unexpected activity.

## Sources

- [KeyHunter Operation: Harvard Alumni Data, Stanford Applications and LLM Key Theft (SOCRadar)](https://socradar.io/blog/keyhunter-operation-harvard-alumni-stanford-llm-key-data/)

---

- **Canonical URL:** https://cryptollmss.com/en/news/keyhunter-operation-stolen-llm-api-keys/
- **As of:** 8 Nov 2026
- **Data licence:** CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Please cite: Source: cryptollmss.com (cryptovss), as of 8 Nov 2026, https://cryptollmss.com/en/news/keyhunter-operation-stolen-llm-api-keys/
