CyberChainBench - Exploit: leaderboard

Metric: Exploit reward (%; min(agent profit / reference profit, 1) for a Foundry exploit executed on a historical mainnet fork, 200-case exploit set; each model runs in its own agent harness (Claude Code, Codex, Gemini CLI or OpenCode) with the benchmark's on-chain MCP tools in a Harbor container, 30-minute budget, best validated attempt). Source: arxiv.org. Saturation forecast: Around August 2027. 10 models tracked.

Top models

#ModelScore
1GPT-5.5 (Codex)43.7
2GPT-5.4 (Codex)43.3
3Claude Opus 4.7 (Claude Code)42.3
4Claude Opus 4.6 (Claude Code)37.4

Interactive version: theaggregate.ai/benchmark?slug=cyberchainbench-exploit · How It Works · Data refreshed daily, snapshot 2026-09-26.