# RULER at 128K context (RULER (128K))

> Synthetic long-context benchmark of retrieval, multi-hop tracing, aggregation and question answering, scored at a 128K-token context.

Canonical page: https://benchlm.ai/benchmarks/ruler128k

- Category: [Reasoning](/reasoning)
- Last updated: October 29, 2026

## About RULER (128K)

- Year: 2024
- Tasks: Synthetic long-context tasks
- Format: Overall accuracy
- Difficulty: Long-context retrieval and tracing
- Paper: [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://github.com/NVIDIA/RULER)

Liquid AI reports it for LFM3-1.7B (October 27, 2026), served with vLLM at full context. Kept apart from the 64K RULER lane because scores fall as context grows.

RULER (128K) is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (1 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [LFM3-1.7B](/models/lfm3-1-7b) | LiquidAI | 65.56% |

## FAQ

### What does RULER (128K) measure?

Synthetic long-context benchmark of retrieval, multi-hop tracing, aggregation and question answering, scored at a 128K-token context.

### Which model scores highest on RULER (128K)?

LFM3-1.7B by LiquidAI currently leads with a score of 65.56% on RULER (128K).

### How many models are evaluated on RULER (128K)?

1 AI models have published results on RULER (128K) in the BenchLM catalog.

### Does RULER (128K) affect BenchLM's overall score?

Not directly. RULER (128K) is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.