# Long-context retrieval and reasoning on RULER 16k context 3k budget

Source: [Long-context retrieval and reasoning on RULER 16k context 3k budget](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-23T11:14:20.117Z

Evaluation of model performance on the RULER dataset configured with a 16k context window and a 3k budget, focusing on long-context retrieval and complex reasoning capabilities.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Long-context retrieval and reasoning |
| Dataset (including any stated split) | RULER 16k context 3k budget |
| Metric | S-1 Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [S-1 Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=S-1%20Score>): Higher is better; unit: PERCENT.

- [S-2 Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=S-2%20Score>): Higher is better; unit: PERCENT.

- [S-3 Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=S-3%20Score>): Higher is better; unit: PERCENT.

- [Key-1 Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Key-1%20Score>): Higher is better; unit: PERCENT.

- [Key-2 Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Key-2%20Score>): Higher is better; unit: PERCENT.

- [Key-3 Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Key-3%20Score>): Higher is better; unit: PERCENT.

- [Validation Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Validation%20Score>): Higher is better; unit: PERCENT.

- [Query Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Query%20Score>): Higher is better; unit: PERCENT.

- [Verification Task Score (VT)](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Verification%20Task%20Score%20(VT)>): Higher is better; unit: PERCENT.

- [Frequency Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Frequency%20Score>): Higher is better; unit: PERCENT.

- [QA Score (1)](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=QA%20Score%20(1)>): Higher is better; unit: PERCENT.

- [QA Score (2)](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=QA%20Score%20(2)>): Higher is better; unit: PERCENT.

- [Average Score](<https://www.sota2.com/research/sota/long-context-retrieval-and-reasoning-on-ruler-16k-context-3k-budget.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | S-1 Score | Result date | Report paper |
| --- | --- | --- | --- |
| LLMLingua | 100 | Not recorded | 69c0c8d70636e5bf4afc394a |
| LongLLMLingua | 100 | Not recorded | 69c0c8d70636e5bf4afc394a |
| BEAVER | 100 | Not recorded | 69c0c8d70636e5bf4afc394a |
| LLMLingua-2 | 27 | Not recorded | 69c0c8d70636e5bf4afc394a |

### LLMLingua

S-1 Score: 100. Result ID: 69c1208247f34ce924a97e77.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c8d70636e5bf4afc394a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71621e209131cea5103a.md>)

Recorded evaluation configuration: {"Avg Tokens":"3,230"}. This record does not assert independent reproduction.

### LongLLMLingua

S-1 Score: 100. Result ID: 69c1208247f34ce924a97e78.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c8d70636e5bf4afc394a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c6d.md>)

[Implementation](<https://aka.ms/LLMLingua>)

Recorded evaluation configuration: {"Avg Tokens":"3,179"}. This record does not assert independent reproduction.

### BEAVER

S-1 Score: 100. Result ID: 69c1208247f34ce924a97e7a.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c8d70636e5bf4afc394a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c0c8d70636e5bf4afc394a.md>)

[Implementation](<https://github.com/JusperLee/BEAVER>)

Recorded evaluation configuration: {"Avg Tokens":"3,198"}. This record does not assert independent reproduction.

### LLMLingua-2

S-1 Score: 27. Result ID: 69c1208247f34ce924a97e79.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c8d70636e5bf4afc394a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61cb2.md>)

[Implementation](<https://aka.ms/LLMLingua-2>)

Recorded evaluation configuration: {"Avg Tokens":"3,132"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.