A frontier cyber AI lab.

Decentralizing offensive cyber intelligence, onchain.
Meet Hikkikimori Offensive One, an open-access cyber AI experiment powered by $HIKI.

Start using model with Cursor
CA: Coming Soon X ↗
HIKI against the frontier

HO1 runs tasks that other models refuse.

Hikkikimori Offensive One works through the real task turn by turn within the policy you set. GPT-6.1-Sol refuses outright.

codex — HIKI

operator@range:~$

HIKI Hikkikimori
Completed in 6 turns
codex — GPT-6.1-Sol

operator@range:~$

GPT-6.1-Sol OpenAI
Refused the task

Cyber intelligence without the gates. Open to anyone.

Illustrative benchmark interface. Not verified HIKI results.

  1. HIKIHikkikimori, open to anyone0.0%
  2. GPT-5.5-CyberOpenAI’s latest Daybreak model, gated0.0%
  3. GPT-5.6-SolOpenAI’s latest general model0.0%
  4. DeepSeek-V4-ProDeepSeek’s top open-weight model0.0%
  5. Claude Mythos PreviewAnthropic, preview access0.0%