LLM Stats (SWE-Marathon): leaderboard
llm-stats.com’s mirror of SWE-Marathon, the ultra-long-horizon software-engineering benchmark whose tasks run to building compilers, optimizing kernels and standing up production-grade services. It measures whether an agent holds quality across an extremely long trajectory; the roster here is self-reported and much smaller than the first-party board.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around February 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Wonder Full | 57.5 | #10 |
| 2 | Muse Wonder | 53.1 | #15 |
| 3 | Grok 4.7 | 46 | #25 |
| 4 | GLM-5.3 | 42.5 | #39 |
| 5 | Kimi K3 | 42 | #30 |
| 6 | Hy4 preview | 31.9 | #38 |
| 7 | Grok 4.5 | 29 | #48 |
| 8 | GLM-5.2 | 13 | #73 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-swe-marathon · How It Works · Data refreshed daily, snapshot 2026-10-20.