PAST-Bench - Self-Improvement Gain: leaderboard
Metric: Gain from retained experience (points on a -100 to 100 scale; overall task score with persistence on minus the matched persistence-off run with the same prompt, grader, tools and seed, averaged per family, then per capability, then over the four capabilities; 26 scenarios and 204 episodes run as ordered sequences of fresh-session tasks; each evaluation episode scores completion (audit-data checks, or a MiniMax-M2.7 rubric judge for open-ended families), tool-error recovery and a safety gate; family-balanced means over three trials; Hermes agent framework). Source: arxiv.org. Saturation forecast: Around January 2028. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 (Medium) | 24 |
| 2 | Claude Sonnet 4.6 | 20 |
| 3 | GLM-5.1 | 20 |
| 4 | Claude Opus 4.6 (Medium) | 19 |
| 5 | Kimi K2.6 (Thinking) | 17 |
| 6 | DeepSeek V4 Pro (High) | 17 |
| 7 | MiniMax-M2.7 | 13 |
Interactive version: theaggregate.ai/benchmark?slug=past-bench-self-improvement-gain · How It Works · Data refreshed daily, snapshot 2026-09-29.