Ergebnisse der KI-Juroren
Vorläufige automatisierte Bewertungen, getrennt von menschlichen Stimmen.
77 validierte Duellurteile
| Rang | Modell | Siege | Niederlagen | Unentschieden | Vergleiche |
|---|---|---|---|---|---|
| 1 | Qwen 3.7 Max | 16 | 4 | 2 | 22 |
| 2 | DeepSeek V4 Pro | 14 | 4 | 1 | 19 |
| 3 | MiMo V2.5 Pro | 14 | 6 | 2 | 22 |
| 4 | Qwen 3.7 Plus | 13 | 9 | 0 | 22 |
| 5 | GLM 5.2 | 9 | 12 | 1 | 22 |
| 6 | DeepSeek V4 Flash | 6 | 15 | 0 | 21 |
| 7 | Gemma 4 31B Instruct QAT | 1 | 0 | 1 | 2 |
| 8 | Qwen 3.6 27B Q4_K_M | 0 | 1 | 1 | 2 |
| 9 | MiMo V2.5 | 0 | 22 | 0 | 22 |
110 validierte Duellurteile
| Rang | Modell | Siege | Niederlagen | Unentschieden | Vergleiche |
|---|---|---|---|---|---|
| 1 | Gemma 4 26B-A4B Instruct QAT | 28 | 19 | 1 | 48 |
| 2 | Qwen 3.6 35B-A3B Q4_K_M | 28 | 20 | 0 | 48 |
| 3 | Gemma 4 31B Instruct QAT | 11 | 19 | 0 | 30 |
| 4 | Qwen 3.6 27B Q4_K_M | 10 | 19 | 1 | 30 |
| 5 | Gemma 4 31B Instruct Q8_0 | 9 | 3 | 0 | 12 |
| 6 | DeepSeek V4 Pro | 6 | 0 | 0 | 6 |
| 7 | MiMo V2.5 Pro | 4 | 2 | 0 | 6 |
| 8 | Qwen 3.7 Max | 3 | 2 | 0 | 5 |
| 9 | Gemma 4 26B-A4B Instruct Q8_0 | 3 | 9 | 0 | 12 |
| 10 | Qwen 3.7 Plus | 2 | 2 | 1 | 5 |
| 11 | DeepSeek V4 Flash | 2 | 3 | 1 | 6 |
| 12 | GLM 5.2 | 2 | 4 | 0 | 6 |
| 13 | MiMo V2.5 | 0 | 6 | 0 | 6 |
337 validierte Duellurteile
| Rang | Modell | Siege | Niederlagen | Unentschieden | Vergleiche |
|---|---|---|---|---|---|
| 1 | Qwen 3.7 Max | 43 | 23 | 0 | 66 |
| 2 | DeepSeek V4 Pro | 43 | 24 | 0 | 67 |
| 3 | Qwen 3.7 Plus | 40 | 26 | 0 | 66 |
| 4 | GLM 5.2 | 39 | 31 | 0 | 70 |
| 5 | MiMo V2.5 Pro | 38 | 29 | 0 | 67 |
| 6 | Gemma 4 26B-A4B Instruct QAT | 34 | 19 | 0 | 53 |
| 7 | DeepSeek V4 Flash | 32 | 34 | 0 | 66 |
| 8 | Gemma 4 31B Instruct QAT | 21 | 13 | 0 | 34 |
| 9 | Qwen 3.6 35B-A3B Q4_K_M | 17 | 36 | 0 | 53 |
| 10 | Qwen 3.6 27B Q4_K_M | 12 | 22 | 0 | 34 |
| 11 | Gemma 4 26B-A4B Instruct Q8_0 | 9 | 5 | 0 | 14 |
| 12 | Gemma 4 31B Instruct Q8_0 | 7 | 7 | 0 | 14 |
| 13 | MiMo V2.5 | 2 | 68 | 0 | 70 |