VibeMemBench (Frozen Verified Experience): leaderboard
Metric: Resolved (%; the frozen verified experience for each target, distilled from earlier trajectories in the same repository and verified to help a reference solver, injected into the solver context; share of 444 target-seed runs (111 SWE-style repository coding targets from 90 SWE-rebench V2 repositories x 4 seeds) whose declared tests pass; MiniSWEAgent solver). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GLM-5.2 (Mini-SWE-Agent) | 80.4 |
| 2 | Qwen3.8-Max (Mini-SWE-Agent) | 80.2 |
| 3 | Kimi K2.7 Code (Mini-SWE-Agent) | 74.3 |
| 4 | DeepSeek-V4-Pro (Mini-SWE-Agent) | 67.1 |
| 5 | GLM-5 (Mini-SWE-Agent) | 57.4 |
Interactive version: theaggregate.ai/benchmark?slug=vibemembench-frozen-verified-experience · How It Works · Data refreshed daily, snapshot 2026-09-26.