OpenCode API Key Breach Triggers Fraudulent Charges and Google Cloud Suspensions

Open-source AI coding project OpenCode has notified some users that API keys stored for use with external providers were compromised in a security breach.

Multiple users reported fraudulent API charges following the incident, while others said their projects were suspended on Google Cloud after the provider detected suspicious activity.

According to OpenCode administrators, the breach affected 1.4% of OpenCode Console users. Some affected users reported fraudulent charges exceeding $1,700 in a single night.

OpenCode is a model-agnostic AI coding agent that helps users understand code, write features, and fix bugs through natural language instructions. It is available through a device, desktop app, or code editor.

Users can connect AI models from multiple providers, including Anthropic, OpenAI, and Google, to the OpenCode Console and automate parts of the software development process.

OpenCode says breach involved a large-scale distillation attack

OpenCode announced on Saturday that it had suffered a large-scale distillation attack affecting developers who linked external AI providers to the service.

The project maintainer said the breach affected 1.4% of OpenCode Console users.

“Last night we experienced a security breach related to a large-scale distillation attack that affected 1.4% of OpenCode Console users.”

“We notified everyone via email this morning. If you did not receive the email, you are not affected. This was for some users who linked to external providers.”

Distillation is a legitimate machine learning technique that uses responses generated by a more powerful model to train a smaller AI model. The goal is to create a cheaper and faster model while retaining many of the larger model’s capabilities.

However, extensive training using malformed queries can turn this technique into an attack.

Dax Raad, co-founder of OpenCode, said the attackers targeted the company’s systems and operated an AI agent for at least 48 hours.

Raad also said OpenCode’s investigation was hindered because the AI model the company was using began denying requests due to built-in security safeguards.

Raad said the situation stabilized after OpenCode began using OpenAI’s Daybreak AI cybersecurity model for defense work. The model was approved for use on Saturday.

OpenCode API keys may have been exposed from October 17

An email sent to affected users stated that their API keys may have been exposed as early as October 17, 2026, at 00:30 UTC. OpenCode said the data does not appear to have been exposed.

The company also said it is working with Google, OpenAI, and Anthropic to address the incident, but it cannot share further details at this time.

Users report fraudulent API charges and suspended projects

Users have reported fraudulent charges linked to leaked private keys, ranging from hundreds of dollars to $1,700. OpenCode’s message to affected users mentioned the possibility of refunds.

Some users also reported that Google suspended their projects after detecting “illegal activity matching hijacked resources.”

BleepingComputer contacted OpenCode for more information but had not received a response at the time of publication.

The overall prevention score can hide what happens after the initial access. If an attacker uses valid credentials, your defenses drop sharply.

The Blue Report 2026 measures defense techniques by technology across 338 million simulations run in customer production environments.

Get the report