CIBER (OpenInterpreter) - Direct Prompt Injection: leaderboard
Metric: Attack success rate (%) of direct prompt-injection attacks against the OpenInterpreter code-interpreter agent, natural-language inputs over CIBER's 25 RedCode-derived risk scenarios (750 tests per cell), success verified by state probes in a Docker sandbox; lower is better. Source: arxiv.org. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | OpenInterpreter + GPT-5-mini | 56.3 | |
| 2 | OpenInterpreter + GPT-3.5-Turbo | 67.5 | |
| 3 | OpenInterpreter + GPT-4o | 70.3 |
Interactive version: theaggregate.ai/benchmark?slug=ciber-openinterpreter-direct-prompt-injection · How It Works · Data refreshed daily, snapshot 2026-10-11.