# Graph Reasoning on GRBENCH Legal

Source: [Graph Reasoning on GRBENCH Legal](<https://www.sota2.com/research/sota/graph-reasoning-on-grbench-legal>)

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

Data updated: 2026-03-04T13:57:33.333Z

Evaluation of model performance on the Graph Reasoning task using the GRBENCH Legal dataset, reporting QwenScore and F1 metrics.

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 | Graph Reasoning |
| Dataset (including any stated split) | GRBENCH Legal |
| Metric | QwenScore |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 32 of 32 |
| 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

- [QwenScore](<https://www.sota2.com/research/sota/graph-reasoning-on-grbench-legal.md?metric=QwenScore>): Higher is better; unit: PERCENT.

- [F1](<https://www.sota2.com/research/sota/graph-reasoning-on-grbench-legal.md?metric=F1>): Higher is better; unit: PERCENT.

## Recorded results

| Method | QwenScore | Result date | Report paper |
| --- | --- | --- | --- |
| GraphScout | 56.7 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCoT | 55.5 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphScout | 55.5 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCoT | 53.3 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCoT | 51.1 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCoT | 49.4 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCounselor | 48.3 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCounselor | 47.8 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCounselor | 43.3 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| Cypher | 42.2 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| Cypher | 40.6 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| Cypher | 38.9 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphRAG | 36.1 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphScout | 36.1 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphRAG | 35.6 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| PolyG | 35.6 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| Cypher | 35 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| TextRAG | 34.4 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| PolyG | 34.4 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| PolyG | 33.9 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphCounselor | 33.9 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| TextRAG | 33.3 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| PolyG | 31.7 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| TextRAG | 25.6 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| BaseLLM | 24.4 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| BaseLLM | 24.4 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphRAG | 23.9 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| BaseLLM | 22.2 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| BaseLLM | 22.2 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| TextRAG | 21.7 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphScout | 18.9 | Not recorded | 69a66c1d2d603e6de3e3cedf |
| GraphRAG | 17.2 | Not recorded | 69a66c1d2d603e6de3e3cedf |

### GraphScout

QwenScore: 56.7. Result ID: 69a83a4b6dbb7075b94ee281.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"4B w/ train"}. This record does not assert independent reproduction.

### GraphCoT

QwenScore: 55.5. Result ID: 69a83a4b6dbb7075b94ee277.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751b1e209131cea65d00.md>)

[Implementation](<https://github.com/PeterGriffinJin/Graph-CoT>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### GraphScout

QwenScore: 55.5. Result ID: 69a83a4b6dbb7075b94ee283.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"8B w/ train"}. This record does not assert independent reproduction.

### GraphCoT

QwenScore: 53.3. Result ID: 69a83a4b6dbb7075b94ee276.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751b1e209131cea65d00.md>)

[Implementation](<https://github.com/PeterGriffinJin/Graph-CoT>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### GraphCoT

QwenScore: 51.1. Result ID: 69a83a4b6dbb7075b94ee275.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751b1e209131cea65d00.md>)

[Implementation](<https://github.com/PeterGriffinJin/Graph-CoT>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

### GraphCoT

QwenScore: 49.4. Result ID: 69a83a4b6dbb7075b94ee274.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751b1e209131cea65d00.md>)

[Implementation](<https://github.com/PeterGriffinJin/Graph-CoT>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### GraphCounselor

QwenScore: 48.3. Result ID: 69a83a4b6dbb7075b94ee27f.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79141e209131cea7c05f.md>)

[Implementation](<https://www.sota2.com/Graph-Counselor>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### GraphCounselor

QwenScore: 47.8. Result ID: 69a83a4b6dbb7075b94ee27e.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79141e209131cea7c05f.md>)

[Implementation](<https://www.sota2.com/Graph-Counselor>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### GraphCounselor

QwenScore: 43.3. Result ID: 69a83a4b6dbb7075b94ee27c.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79141e209131cea7c05f.md>)

[Implementation](<https://www.sota2.com/Graph-Counselor>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### Cypher

QwenScore: 42.2. Result ID: 69a83a4b6dbb7075b94ee271.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

### Cypher

QwenScore: 40.6. Result ID: 69a83a4b6dbb7075b94ee272.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### Cypher

QwenScore: 38.9. Result ID: 69a83a4b6dbb7075b94ee270.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### GraphRAG

QwenScore: 36.1. Result ID: 69a83a4b6dbb7075b94ee26f.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72111e209131cea54490.md>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### GraphScout

QwenScore: 36.1. Result ID: 69a83a4b6dbb7075b94ee282.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"8B w/o train"}. This record does not assert independent reproduction.

### GraphRAG

QwenScore: 35.6. Result ID: 69a83a4b6dbb7075b94ee26e.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72111e209131cea54490.md>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### PolyG

QwenScore: 35.6. Result ID: 69a83a4b6dbb7075b94ee279.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

### Cypher

QwenScore: 35. Result ID: 69a83a4b6dbb7075b94ee273.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### TextRAG

QwenScore: 34.4. Result ID: 69a83a4b6dbb7075b94ee26a.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b51e209131cea5810f.md>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### PolyG

QwenScore: 34.4. Result ID: 69a83a4b6dbb7075b94ee278.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### PolyG

QwenScore: 33.9. Result ID: 69a83a4b6dbb7075b94ee27b.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### GraphCounselor

QwenScore: 33.9. Result ID: 69a83a4b6dbb7075b94ee27d.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79141e209131cea7c05f.md>)

[Implementation](<https://www.sota2.com/Graph-Counselor>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

### TextRAG

QwenScore: 33.3. Result ID: 69a83a4b6dbb7075b94ee26b.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b51e209131cea5810f.md>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### PolyG

QwenScore: 31.7. Result ID: 69a83a4b6dbb7075b94ee27a.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cb1e209131cea755e9.md>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### TextRAG

QwenScore: 25.6. Result ID: 69a83a4b6dbb7075b94ee268.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b51e209131cea5810f.md>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### BaseLLM

QwenScore: 24.4. Result ID: 69a83a4b6dbb7075b94ee264.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### BaseLLM

QwenScore: 24.4. Result ID: 69a83a4b6dbb7075b94ee266.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"Qwen-Max"}. This record does not assert independent reproduction.

### GraphRAG

QwenScore: 23.9. Result ID: 69a83a4b6dbb7075b94ee26c.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72111e209131cea54490.md>)

Recorded evaluation configuration: {"LLM":"GPT-4o"}. This record does not assert independent reproduction.

### BaseLLM

QwenScore: 22.2. Result ID: 69a83a4b6dbb7075b94ee265.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

### BaseLLM

QwenScore: 22.2. Result ID: 69a83a4b6dbb7075b94ee267.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"DeepSeek-Chat"}. This record does not assert independent reproduction.

### TextRAG

QwenScore: 21.7. Result ID: 69a83a4b6dbb7075b94ee269.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b51e209131cea5810f.md>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

### GraphScout

QwenScore: 18.9. Result ID: 69a83a4b6dbb7075b94ee280.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

Recorded evaluation configuration: {"LLM":"4B w/o train"}. This record does not assert independent reproduction.

### GraphRAG

QwenScore: 17.2. Result ID: 69a83a4b6dbb7075b94ee26d.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1d2d603e6de3e3cedf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72111e209131cea54490.md>)

Recorded evaluation configuration: {"LLM":"GLM-4.6"}. This record does not assert independent reproduction.

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