LongDS v1.1 Lite: leaderboard
Metric: Accuracy (%; judged turn accuracy macro-averaged over 24 tasks), each model under one agentic coding harness. Source: zjunlp.github.io. Saturation forecast: Around December 2026. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 | 78.17 |
| 2 | Claude Fable 5.1 | 76.53 |
| 3 | GPT-5.6 Sol | 70.71 |
| 4 | Kimi K3 | 68.9 |
| 5 | Qwen 3.8 Max (0902) | 62.58 |
| 6 | GLM-5.2 | 56.59 |
| 7 | DeepSeek V4 Pro | 39.33 |
Interactive version: theaggregate.ai/benchmark?slug=longds-v1-1-lite · How It Works · Data refreshed daily, snapshot 2026-09-25.