Kegagalan kategori AI BENCHY
Trik anti-AI: Format tambahan
Trik anti-AI
Format tambahan
Lihat model AI mana yang paling mungkin mengalami Format tambahan di Trik anti-AI, agar Anda bisa menemukan titik lemahnya lebih cepat.
Alasan kegagalan
| Peringkat | Model | Perusahaan | Jumlah Format tambahan | Skor kategori | Tes benar | Waktu respons (rata-rata) |
|---|---|---|---|---|---|---|
| #68 | Claude Opus 4.8 none | Anthropic | 2 | 6.5 | 2/4 | 3.40s |
| #69 | Claude Opus 4.6 medium | Anthropic | 2 | 6.4 | 2/4 | 7.45s |
| #77 | Claude Sonnet 4.6 none | Anthropic | 2 | 4.8 | 1/4 | 2.94s |
| #133 | DeepSeek V3.2 none | DeepSeek | 2 | 3.2 | 0/4 | 9.35s |
| #30 | Qwen3.5-27B medium | Qwen | 1 | 8.7 | 3/4 | 19.8s |
| #47 | Grok Build 0.1 medium | X AI | 1 | 8.3 | 3/4 | 7.43s |
| #52 | Claude Sonnet 4.6 medium | Anthropic | 1 | 6.5 | 2/4 | 2.98s |
| #64 | MiMo-V2-Flash medium | Xiaomi | 1 | 8.1 | 3/4 | 15.8s |
| #84 | Grok 4.20 Multi Agent Beta medium | X AI | 1 | 6.9 | 2/4 | 3.46s |
| #140 | Qwen3 Coder Next none | Qwen | 1 | 3.6 | 0/4 | 3.31s |