RSI-Exam: New Benchmark Tests AI Agents' Recursive Self-Improvement Across 88 Tasks
yuyinzhou_cs · x · 2026-08-27
As frontier models improve monthly, Recursive Self-Improvement (RSI) is hot — but how to measure it? Researchers introduce RSI-Exam: 88 executable research tasks across 6 domains (AI agents, virtual cells, TPU kernels, chip design, quant finance, distillation, and more), testing whether agents can improve a working method via long-horizon autonomous experimentation and generalize to hidden data.
Protocol: an agent inherits a working method and iterates on visible data for hours; only the final artifact is evaluated once in a fresh container on a hidden set.
Findings: Claude and GPT form a clear first tier (Opus 5 leads the 88-task leaderboard by mean hidden-set score), with a large gap over the rest — yet huge headroom remains. Trajectories show both agents discovering fundamentally better methods and agents spending hours optimizing the wrong idea. Task contributions are open to all fields.
More from Models
- Nat Lambert: wave of AI releases rumored — thinky, muse spark, Fable 5.5, OpenAI — natolambert · 2026-10-15
- Musk amplifies dev's verdict: Grok Bot easier to use than Codex and Claude — elonmusk · 2026-10-15
- Rumor: New Anthropic model imminent, apparently behind a batch of high-severity Chrome CVEs — matthew_d_green · 2026-10-15
- StartLux-Decision debuts RL-trained calibrated decision models, 27B scores 72.8 — startlux-models · 2026-10-15
- Meta's Muse Code slammed for harsh quotas: 21% left after 5 hours — ThomasScialom · 2026-10-15
- Claude Code error mentions 'Fable 5.1' and 'Opus 5', hinting at unreleased models — adonis_singh · 2026-10-15