# Multi-turn Coding on Novel 35-turn coding workflow

Source: [Multi-turn Coding on Novel 35-turn coding workflow](<https://www.sota2.com/research/sota/multi-turn-coding-on-novel-35-turn-coding-workflow>)

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

Data updated: 2026-05-27T08:00:27.701Z

Evaluation of model performance on multi-turn coding workflows, focusing on latency metrics such as median time per turn and total wall time.

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 | Multi-turn Coding |
| Dataset (including any stated split) | Novel 35-turn coding workflow |
| Metric | Median Time per Turn (ms) |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 3 of 3 |
| 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

- [Median Time per Turn (ms)](<https://www.sota2.com/research/sota/multi-turn-coding-on-novel-35-turn-coding-workflow.md?metric=Median%20Time%20per%20Turn%20(ms)>): Lower is better; unit: TIME.

- [Total Wall Time (s)](<https://www.sota2.com/research/sota/multi-turn-coding-on-novel-35-turn-coding-workflow.md?metric=Total%20Wall%20Time%20(s)>): Lower is better; unit: TIME.

- [Relative Wall Time vs LayerScale](<https://www.sota2.com/research/sota/multi-turn-coding-on-novel-35-turn-coding-workflow.md?metric=Relative%20Wall%20Time%20vs%20LayerScale>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Median Time per Turn (ms) | Result date | Report paper |
| --- | --- | --- | --- |
| LayerScale | 647 | Not recorded | 6a167aa3dbd7150506f8002d |
| SGLang | 2732 | Not recorded | 6a167aa3dbd7150506f8002d |
| vLLM | 2743 | Not recorded | 6a167aa3dbd7150506f8002d |

### LayerScale

Median Time per Turn (ms): 647. Result ID: 6a16a3e2fbc373592c0f9319.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa3dbd7150506f8002d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a167aa3dbd7150506f8002d.md>)

Recorded evaluation configuration: {"caching":"no response-cache hits"}. This record does not assert independent reproduction.

### SGLang

Median Time per Turn (ms): 2732. Result ID: 6a16a3e2fbc373592c0f931b.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa3dbd7150506f8002d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738a1e209131cea5d174.md>)

Recorded evaluation configuration: {"caching":"no response-cache hits"}. This record does not assert independent reproduction.

### vLLM

Median Time per Turn (ms): 2743. Result ID: 6a16a3e2fbc373592c0f931a.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa3dbd7150506f8002d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa708c1e209131cea4ba94.md>)

Recorded evaluation configuration: {"caching":"no response-cache hits"}. This record does not assert independent reproduction.

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