Catégorie AI BENCHY
Classement Combiné
Voyez quels modèles d'IA réussissent le mieux sur Combiné, lesquels restent fiables et où les écarts sont les plus marqués.
| Rang | Modèle | Entreprise | Score Combiné | Score | Tests corrects | Temps de réponse (moy.) |
|---|---|---|---|---|---|---|
| #47 | Grok 4.20 medium | X AI | 10.0 | 7.0 | 1/1 | 17.4s |
| #51 | Nemotron 3 Super medium | NVIDIA | 10.0 | 6.7 | 1/1 | 87.8s |
| #52 | Grok 4.1 Fast medium | X AI | 10.0 | 6.7 | 1/1 | 37.6s |
| #54 | Mercury 2 medium | Inception | 10.0 | 6.5 | 1/1 | 3.28s |
| #57 | GPT-5 Nano medium | OpenAI | 10.0 | 6.3 | 1/1 | 66.0s |
| #68 | gpt-oss-120b medium | OpenAI | 10.0 | 5.8 | 1/1 | 31.2s |
| #38 | GPT-5.4 Nano medium | OpenAI | 9.8 | 7.6 | 1/1 | 24.1s |
| #41 | MiMo-V2-Flash medium | Xiaomi | 9.8 | 7.5 | 1/1 | 75.7s |
| #24 | Gemma 4 26B A4B medium | 9.6 | 8.0 | 1/1 | 73.5s | |
| #2 | Gemini 3.1 Pro Preview medium | 9.5 | 9.6 | 1/1 | 40.6s | |
| #4 | Claude Opus 4.7 none | Anthropic | 9.5 | 9.2 | 1/1 | 18.3s |
| #33 | GLM 5.1 medium | Z.ai | 9.5 | 7.8 | 1/1 | 43.1s |
| #42 | Claude Sonnet 4.6 none | Anthropic | 9.5 | 7.4 | 1/1 | 23.8s |
| #31 | GLM 5V Turbo medium | Z.ai | 6.9 | 7.8 | 0/1 | 15.1s |
| #64 | DeepSeek V3.2 none | DeepSeek | 6.5 | 6.1 | 0/1 | 115.9s |