SEQUOR Replace 10: leaderboard
Metric: Replace constraints every 10 turns: last-minus-first-turn accuracy change (percentage points) (self-reported). Source: benchmarklist.com. Saturation forecast: Around 2030. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GLM-4.7 Flash | -9.26 |
| 2 | Gemini 3.1 Flash Lite | -11.5 |
| 3 | Gemma 3 12B | -14.5 |
| 4 | Llama 3.3 70B Instruct | -18 |
| 5 | Gemma 3 27B | -19 |
| 6 | GPT-OSS-20B | -35.5 |
| 7 | Gemma 3 4B | -44.5 |
| 8 | GPT-OSS-120B | -45.5 |
Interactive version: theaggregate.ai/benchmark?slug=sequor-replace-10 · How It Works · Data refreshed daily, snapshot 2026-10-09.