EdgeBench - Science (Claude Code 1M): leaderboard
Metric: Scientific discovery and machine learning tasks: mean best-so-far score (0-100) after a 12-hour continuous agent run with evaluator feedback, nominally three trials per task with only valid trajectories scored, averaged over the family's tasks; higher is better; harness/context: Claude Code 1M. Source: arxiv.org. 1 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 (Claude Code) | 48.5 |
Interactive version: theaggregate.ai/benchmark?slug=edgebench-science-claude-code-1m · How It Works · Data refreshed daily, snapshot 2026-10-09.