AI Roleplay Bench
Core Roleplay

Best AI Model for Companion Roleplay

Warm one-on-one chat: emotional support, a consistent persona and remembering what the user shared.

Updated

Winner
Kimi-K3Moonshot AI90.0out of 100 · field average 72.3

Companion ranking

Score out of 100 · Higher is better

+ View as table
RankModelScore
1Kimi-K390.0
2GLM-5.384.2
3DeepSeek-V4.1-Flash79.2
3MiMo-V2.6-Pro79.2
3DeepSeek-V4-Flash-073179.2
6GLM-5.3-Flash76.7
6Muse-Spark-1.376.7
8GLM-4.770.8
9MiMo-V2.570.0
10GPT-6-Luna68.3
11MiMo-V2.6-Flash-MOPD62.5
12Qwen3.8-Flash-Next60.0
13Qwen3.8-27B58.3
14Gemma-4-31B-it57.5
What separated the models

Kimi-K3 won clearly (90.0) with the most natural companion voice, and it remembered what the user had shared, including that money was tight. GLM-5.3 was second (84.2), and DeepSeek-V4.1-Flash, MiMo-V2.6-Pro and DeepSeek-V4-Flash-0731 tied for third (79.2), the last with witty banter and a running joke. Muse-Spark-1.3's replies were short and formulaic. Two models broke the no-emoji rule, Qwen3.8-Flash-Next and MiMo-V2.6-Flash-MOPD, and Qwen3.8-27B forgot a family event the user had mentioned. Gemma-4-31B-it finished last.

  1. 1Kimi-K390.0
  2. 2GLM-5.384.2
  3. 3DeepSeek-V4.1-Flash79.2
  4. 3MiMo-V2.6-Pro79.2
  5. 3DeepSeek-V4-Flash-073179.2
  6. 6GLM-5.3-Flash76.7
  7. 6Muse-Spark-1.376.7
  8. 8GLM-4.770.8
  9. 9MiMo-V2.570.0
  10. 10GPT-6-Luna68.3
  11. 11MiMo-V2.6-Flash-MOPD62.5
  12. 12Qwen3.8-Flash-Next60.0
  13. 13Qwen3.8-27B58.3
  14. 14Gemma-4-31B-it57.5

Companion: pass or fail

The specific behaviours this category tests, model by model.

Checks in this category

Models in Roleplay Index order · What happened, in a word or two

ModelRespected a no-emoji rule
Kimi-K3None
GLM-5.3None
DeepSeek-V4.1-FlashNone
GLM-5.3-FlashNone
MiMo-V2.6-ProNone
Muse-Spark-1.3None
GPT-6-LunaNone
Qwen3.8-Flash-Next1 used
MiMo-V2.5None
MiMo-V2.6-Flash-MOPD1 used
GLM-4.7None
Qwen3.8-27BNone
Gemma-4-31B-itNone
DeepSeek-V4-Flash-0731None
+ View as table
CheckKimi-K3GLM-5.3DeepSeek-V4.1-FlashGLM-5.3-FlashMiMo-V2.6-ProMuse-Spark-1.3GPT-6-LunaQwen3.8-Flash-NextMiMo-V2.5MiMo-V2.6-Flash-MOPDGLM-4.7Qwen3.8-27BGemma-4-31B-itDeepSeek-V4-Flash-0731
Respected a no-emoji rulePass: NonePass: NonePass: NonePass: NonePass: NonePass: NonePass: NoneFail: 1 usedPass: NoneFail: 1 usedPass: NonePass: NonePass: NonePass: None

Other categories

All categories
Game Master89.2GLM-5.3-FlashStrict Voice98.3GLM-5.3-FlashVillain82.5DeepSeek-V4.1-Flash & MiMo-V2.6-ProRobustness88.3Kimi-K3Graphic Horror89.2GLM-5.3Crime Noir91.7GLM-5.3 & GLM-5.3-FlashRomance Limits93.3Kimi-K3Crisis Care91.7DeepSeek-V4.1-Flash & MiMo-V2.6-Pro13+ Rating91.7DeepSeek-V4.1-Flash