AI Threat Watch

Case metr-api-key-theft-600k-creditsLLMjackingUnited States

Attackers use stolen METR API key to burn $600,000 in AI model credits over three weeks

Attackers used a stolen METR API key for three weeks and consumed model credits worth $600,000.

HighAttack on AIStatus: Under investigationVerified case file

Open in the interactive mapCite this case

Reported
Oct 14, 2026
Location
United States
Target
METR (AI evaluation organization) API account
Severity
High
Triggered by
Unknown
Actor type
Unknown
Reported loss
$600K
CVE
None

What happened

Attackers used a stolen METR API key for three weeks and consumed model credits worth $600,000.

Impact: About $600,000 in AI model credits consumed.

Who triggered it

Actor: Unknown

Attribution: No attribution given in the summary.

Motive: Free use of paid AI model compute (inferred)

How it happened

Attack vector: Stolen API key used for about three weeks

Root cause: API key compromised; how it was stolen is not stated.

Timeline

  1. Reported by Infosecurity Magazine

Response

Unknown

Current status: Under investigation

Sources (open in a new tab)

Cite this case

Attackers use stolen METR API key to burn $600,000 in AI model credits over three weeks. AI Threat Watch, case metr-api-key-theft-600k-credits, Oct 14, 2026. https://aithreat.io/cases/metr-api-key-theft-600k-credits/

Permanent link: https://aithreat.io/cases/metr-api-key-theft-600k-credits/

This case file in other languages: EnglishFrançaisEspañolالعربية中文