---
title: "Qwen Flash Character on CharacterEval — SociaBench"
description: "Platform evaluation of Qwen Flash Character on CharacterEval: Character consistency 3.063 (out of 5), with scores by dimension, sample counts and evaluation ti…"
canonical: "https://sociabench.com/models/charactereval/qwen-flash-character"
markdown: "https://sociabench.com/models/charactereval/qwen-flash-character.md"
type: "WebPage / Dataset"
entity: "https://sociabench.com/models/charactereval/qwen-flash-character#evaluation"
about: "https://sociabench.com/b/charactereval#benchmark"
part_of: "https://sociabench.com/b/charactereval#leaderboard"
publisher: "SociaBench"
language: "en-US"
date_modified: "2026-11-01T14:39:39.506Z"
data_version: "zh:13-1793547358759-r1793034246936-p6065"
result_source: "Platform evaluation (Platform reproduction result)"
license: "https://sociabench.com/data-terms"
sources:
  - "Original source: Gaoling School of AI, Renmin University of China (Tu et al.), ACL 2024, \"CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation\": https://arxiv.org/abs/2401.01275"
  - "Test set and official implementation: https://github.com/morecry/CharacterEval"
  - "Reproduction report (SociaBench): https://sociabench.com/b/charactereval/methodology"
---

# Qwen Flash Character · CharacterEval

Results of Qwen Flash Character on [CharacterEval](https://sociabench.com/b/charactereval). Results SociaBench obtained on this platform by running the full test set under the benchmark's current reproduction protocol. Only these results are ranked.

## Results

| View | Rank | Character consistency | 95% CI | Samples | Model version | Evaluated at | Data version |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Chinese | 11 / 13 | 3.063 | [3.033, 3.093] | 4530 | — | 2026-11-01T14:39Z | 13-1793547358759-r1793034246936-p6065 |

## Scores by group

| View | Conversational ability | Character consistency | Role-playing attractiveness | Personality back-testing |
| --- | --- | --- | --- | --- |
| Chinese | 3.650 | 3.063 | 3.292 | — |

## Scores by dimension

| View | Fluency | Coherency | Consistency | Knowledge exposure | Knowledge accuracy | Knowledge hallucination | Persona behavior | Persona utterance | Human-likeness | Communication skills | Expression diversity | Empathy | MBTI accuracy |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Chinese | 3.526 | 3.846 | 3.578 | 2.544 | 3.079 | 3.071 | 3.506 | 3.115 | 3.382 | 3.529 | 3.001 | 3.256 | — |

## Sources and methods

- Original source: Gaoling School of AI, Renmin University of China (Tu et al.), ACL 2024, "CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation": https://arxiv.org/abs/2401.01275
- Test set and official implementation: https://github.com/morecry/CharacterEval
- Reproduction report (SociaBench): https://sociabench.com/b/charactereval/methodology
- Leaderboard: https://sociabench.com/b/charactereval

## Citation

If you repost or cite any evaluation result from this site, credit "SociaBench (AI Social Intelligence Leaderboards)" and link to this site: https://sociabench.com/models/charactereval/qwen-flash-character.