Kegagalan kategori AI BENCHY
Kepatuhan instruksi: Jawaban salah
Kepatuhan instruksi
Jawaban salah
Lihat model AI mana yang paling mungkin mengalami Jawaban salah di Kepatuhan instruksi, agar Anda bisa menemukan titik lemahnya lebih cepat. Urutkan berdasarkan: Tes benar ↓.
Alasan kegagalan
| Peringkat | Model | Perusahaan | Jumlah Jawaban salah | Skor kategori | Tes benar | Waktu respons (rata-rata) |
|---|---|---|---|---|---|---|
| #21 | Gemini 3 Flash Preview none | 1 | 6.4 | 1/2 | 1.58s | |
| #28 | GPT-5.2 Chat none | OpenAI | 1 | 7.5 | 1/2 | 5.46s |
| #33 | GLM 5.1 medium | Z.ai | 1 | 6.4 | 1/2 | 7.47s |
| #36 | GPT-5.3 Chat none | OpenAI | 1 | 8.3 | 1/2 | 3.29s |
| #42 | Claude Sonnet 4.6 none | Anthropic | 1 | 6.5 | 1/2 | 1.96s |
| #48 | Gemma 4 31B none | 1 | 6.5 | 1/2 | 2.84s | |
| #55 | MiMo-V2-Omni none | Xiaomi | 1 | 6.5 | 1/2 | 4.18s |
| #58 | GLM 5V Turbo none | Z.ai | 1 | 6.5 | 1/2 | 1.97s |
| #59 | Qwen3.5-Flash none | Qwen | 1 | 6.3 | 1/2 | 8.81s |
| #62 | Gemini 2.5 Flash none | 1 | 8.0 | 1/2 | 672ms | |
| #63 | Qwen3.5-35B-A3B none | Qwen | 1 | 6.3 | 1/2 | 809ms |
| #65 | MiMo-V2-Pro none | Xiaomi | 1 | 6.5 | 1/2 | 2.51s |
| #66 | GPT-5.4 none | OpenAI | 1 | 6.5 | 1/2 | 1.07s |
| #69 | Kimi K2.6 none | Moonshot AI | 1 | 6.5 | 1/2 | 1.64s |
| #72 | Hunter Alpha none | OpenRouter | 1 | 6.4 | 1/2 | 2.82s |