Échecs par catégorie AI BENCHY
Résolution d'énigmes
Mauvaise réponse
Résolution d'énigmes
Mauvaise réponse
Voyez quels modèles d'IA ont le plus de chances de rencontrer Mauvaise réponse sur Résolution d'énigmes, pour repérer plus vite les points faibles. Trier par: Nombre d'échecs ↑.
Raisons d'échec liées
| Rang | Modèle | Entreprise | Nombre de Mauvaise réponse | Score de catégorie | Tests corrects | Temps de réponse (moy.) |
|---|---|---|---|---|---|---|
| #8 | Gemini 3.1 Flash Lite Preview high | 1 | 7.0 | 2/3 | 46.3s | |
| #12 | Gemini 3.1 Flash Lite Preview medium | 1 | 7.0 | 2/3 | 3.58s | |
| #13 | Step 3.5 Flash medium | Stepfun | 1 | 4.0 | 1/3 | 7.72s |
| #15 | GPT-5.2 Chat none | OpenAI | 1 | 7.0 | 2/3 | 4.42s |
| #16 | Gemini 2.5 Flash medium | 1 | 7.0 | 2/3 | 3.94s | |
| #18 | DeepSeek V3.2 medium | DeepSeek | 1 | 7.0 | 2/3 | 36.9s |
| #20 | Gemini 3 Flash Preview none | 1 | 7.0 | 2/3 | 1.06s | |
| #21 | MiMo-V2-Flash medium | Xiaomi | 1 | 7.0 | 2/3 | 3.77s |
| #23 | Seed-2.0-Mini medium | Bytedance Seed | 1 | 7.0 | 2/3 | 25.9s |
| #26 | Claude Opus 4.6 medium | Anthropic | 1 | 7.0 | 2/3 | 4.60s |
| #28 | Kimi K2.5 medium | Moonshot AI | 1 | 4.0 | 1/3 | 45.4s |
| #29 | Qwen3.5 Plus 2026-02-15 none | Qwen | 1 | 7.0 | 2/3 | 2.82s |
| #30 | Grok 4.1 Fast medium | X AI | 1 | 4.0 | 1/3 | 8.08s |
| #31 | GLM 5 none | Z.ai | 1 | 7.0 | 2/3 | 2.05s |
| #32 | GPT-5 Mini medium | OpenAI | 1 | 4.3 | 1/3 | 14.1s |
| #33 | DeepSeek V3.2 none | DeepSeek | 1 | 7.7 | 2/3 | 7.37s |
| #34 | GPT-5 Nano medium | OpenAI | 1 | 4.0 | 1/3 | 19.8s |
| #35 | Qwen3.5-35B-A3B medium | Qwen | 1 | 4.0 | 1/3 | 31.6s |
| #36 | Mercury 2 medium | Inception | 1 | 1.7 | 0/3 | 934ms |
| #38 | Gemini 2.5 Flash none | 1 | 4.7 | 1/3 | 576ms | |
| #39 | gpt-oss-120b medium | OpenAI | 1 | 1.7 | 0/3 | 11.8s |
| #41 | Qwen3.5-27B none | Qwen | 1 | 6.3 | 1/3 | 1.37s |
| #43 | MiniMax M2.5 medium | Minimax | 1 | 4.0 | 1/3 | 11.5s |
| #44 | GPT-5.4 none | OpenAI | 1 | 4.0 | 1/3 | 1.52s |
| #49 | GLM 4.7 Flash none | Z.ai | 1 | 3.7 | 0/3 | 1.00s |
| #50 | Qwen3 Coder Next medium | Qwen | 1 | 10.0 | 0/3 | 2.30s |
| #55 | LFM2-24B-A2B none | Liquid | 1 | 3.3 | 0/3 | 1.69s |
| #37 | Qwen3.5-Flash none | Qwen | 2 | 1.3 | 0/3 | 5.90s |
| #40 | Qwen3.5-122B-A10B none | Qwen | 2 | 4.0 | 1/3 | 982ms |
| #42 | Qwen3.5-35B-A3B none | Qwen | 2 | 1.7 | 0/3 | 1.34s |
| #45 | Trinity Large Preview none | Arcee AI | 2 | 4.0 | 1/3 | 3.30s |
| #52 | GLM 4.7 Flash medium | Z.ai | 2 | 10.0 | 0/3 | 12.9s |
| #46 | Kimi K2.5 none | Moonshot AI | 3 | 10.0 | 0/3 | 4.73s |
| #47 | GPT-4o-mini none | OpenAI | 3 | 2.3 | 0/3 | 1.30s |
| #48 | Qwen3 Coder Next none | Qwen | 3 | 1.3 | 0/3 | 22.9s |
| #51 | Mercury 2 none | Inception | 3 | 10.0 | 0/3 | 533ms |
| #53 | Grok 4.1 Fast none | X AI | 3 | 1.3 | 0/3 | 1.28s |
| #54 | MiMo-V2-Flash none | Xiaomi | 3 | 10.0 | 0/3 | 1.38s |