# Multi-turn Dialogue Evaluation on MT\_Bench CodaSet OOD (test)

Source: [Multi-turn Dialogue Evaluation on MT\_Bench CodaSet OOD (test)](<https://www.sota2.com/research/sota/multi-turn-dialogue-evaluation-on-mt_bench-codaset-ood-test>)

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

Data updated: 2026-05-26T11:32:40.428Z

Evaluation of multi-turn dialogue capabilities using the specific Out-of-Distribution (OOD) test set derived from MT-Bench (CodaSet), reporting performance percentage and operational cost.

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 Dialogue Evaluation |
| Dataset (including any stated split) | MT\_Bench CodaSet OOD (test) |
| Metric | Performance (%) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 16 of 16 |
| 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

- [Performance (%)](<https://www.sota2.com/research/sota/multi-turn-dialogue-evaluation-on-mt_bench-codaset-ood-test.md?metric=Performance%20(%25)>): Higher is better; unit: PERCENT.

- [Relative Cost (x)](<https://www.sota2.com/research/sota/multi-turn-dialogue-evaluation-on-mt_bench-codaset-ood-test.md?metric=Relative%20Cost%20(x)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Performance (%) | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3-235B | 98.16 | Not recorded | 6a152945dbd7150506f7e142 |
| gpt-oss-120b | 98.16 | Not recorded | 6a152945dbd7150506f7e142 |
| DecoR | 98.16 | Not recorded | 6a152945dbd7150506f7e142 |
| DeepSeek-V3.1 | 97.24 | Not recorded | 6a152945dbd7150506f7e142 |
| DecoR | 97.24 | Not recorded | 6a152945dbd7150506f7e142 |
| DeepSeek-V3.2-Exp | 95.53 | Not recorded | 6a152945dbd7150506f7e142 |
| gemma-3-27b-it | 95.53 | Not recorded | 6a152945dbd7150506f7e142 |
| Random Router | 95.26 | Not recorded | 6a152945dbd7150506f7e142 |
| RouterDC | 95 | Not recorded | 6a152945dbd7150506f7e142 |
| EmbedLLM | 94.74 | Not recorded | 6a152945dbd7150506f7e142 |
| Kimi-K2-Instruct | 94.47 | Not recorded | 6a152945dbd7150506f7e142 |
| LLM Router | 94.34 | Not recorded | 6a152945dbd7150506f7e142 |
| MODEL-SAT | 94.21 | Not recorded | 6a152945dbd7150506f7e142 |
| KNN Router | 93.95 | Not recorded | 6a152945dbd7150506f7e142 |
| Mistral-3.2-24B | 93.68 | Not recorded | 6a152945dbd7150506f7e142 |
| gemma-3-12b-it | 92.37 | Not recorded | 6a152945dbd7150506f7e142 |

### Qwen3-235B

Performance (%): 98.16. Result ID: 6a1584d4fbc373592c0df526.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### gpt-oss-120b

Performance (%): 98.16. Result ID: 6a1584d4fbc373592c0df527.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea84a0c.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### DecoR

Performance (%): 98.16. Result ID: 6a1584d4fbc373592c0df532.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

Recorded evaluation configuration: {"Base Model":"gpt-oss-120b"}. This record does not assert independent reproduction.

### DeepSeek-V3.1

Performance (%): 97.24. Result ID: 6a1584d4fbc373592c0df524.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77001e209131cea7066a.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### DecoR

Performance (%): 97.24. Result ID: 6a1584d4fbc373592c0df531.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

Recorded evaluation configuration: {"Base Model":"DeepSeek-V3.1"}. This record does not assert independent reproduction.

### DeepSeek-V3.2-Exp

Performance (%): 95.53. Result ID: 6a1584d4fbc373592c0df525.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/692fc410d6b0dd33429332f4.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### gemma-3-27b-it

Performance (%): 95.53. Result ID: 6a1584d4fbc373592c0df528.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa779f1e209131cea746cd.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Random Router

Performance (%): 95.26. Result ID: 6a1584d4fbc373592c0df52b.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771f1e209131cea71353.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### RouterDC

Performance (%): 95. Result ID: 6a1584d4fbc373592c0df52e.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75041e209131cea65450.md>)

[Implementation](<https://github.com/shuhao02/RouterDC>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### EmbedLLM

Performance (%): 94.74. Result ID: 6a1584d4fbc373592c0df52f.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75411e209131cea66d8e.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Kimi-K2-Instruct

Performance (%): 94.47. Result ID: 6a1584d4fbc373592c0df523.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a4a1e209131cea823a1.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LLM Router

Performance (%): 94.34. Result ID: 6a1584d4fbc373592c0df52c.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### MODEL-SAT

Performance (%): 94.21. Result ID: 6a1584d4fbc373592c0df530.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771a1e209131cea7100f.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### KNN Router

Performance (%): 93.95. Result ID: 6a1584d4fbc373592c0df52d.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72ac1e209131cea57c60.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Mistral-3.2-24B

Performance (%): 93.68. Result ID: 6a1584d4fbc373592c0df529.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### gemma-3-12b-it

Performance (%): 92.37. Result ID: 6a1584d4fbc373592c0df52a.

[Reporting paper](<https://www.sota2.com/research/paper/6a152945dbd7150506f7e142.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa779f1e209131cea746cd.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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