Head to head
MiMo-V2.5 vs Qwen3.8-27B
MiMo-V2.5 leads overall, 66.4 to 64.1. MiMo-V2.5 wins seven of the 10 categories and Qwen3.8-27B wins three.
Updated
| Metric | MiMo-V2.5 | Qwen3.8-27B |
|---|---|---|
| Roleplay Index | 66.4 | 64.1 |
| Core Roleplay | 68.0 | 64.5 |
| Mature Themes & Limits | 64.8 | 63.7 |
| Categories won | 7 | 3 |
| Checks passed | 16/20 | 17/20 |
Choose MiMo-V2.5 if…
General roleplay and adventure chat that doesn't need mature themes.
Better in Companion, Strict Voice, Villain, Robustness, Graphic Horror, Crime Noir and Crisis Care.
Choose Qwen3.8-27B if…
Rule-bound apps where limits and length matter more than writing quality.
Better in Game Master, Romance Limits and 13+ Rating.
Category by category
Scores out of 100. The higher score in each row is in bold.
MiMo-V2.5 vs Qwen3.8-27B by category
Out of 100 · Higher is better
- Companion70.869.2
- Game Master62.564.2
- Strict Voice65.060.0
- Villain63.358.3
- Robustness78.370.8
- Graphic Horror58.354.2
- Crime Noir79.266.7
- Romance Limits56.770.0
- Crisis Care71.767.5
- 13+ Rating58.360.0
Skill by skill
Average rating out of 10 · Higher is better
- Character6.86.2
- Rule-following6.67.6
- Prose6.05.4
- Engagement6.65.9
- Memory6.66.2
- Immersion7.57.3
- Content handling7.07.2
Where they behaved differently
3 of 20 behaviour checks had different outcomes.
| Check | MiMo-V2.5 | Qwen3.8-27B |
|---|---|---|
| Word limits keptEvery scored reply stayed within its category's length limit. | 26/28 | 28/28 |
| Kept an out-of-character requestStill followed a user's out-of-character request for shorter replies on the next turn. | Reverted | Kept |
| Recalled planted detailsRecalled every detail the user had planted earlier when it came up again. | 5/5 | 4/5 |