Categorie AI BENCHY
Clasament Programare
Vezi ce modele AI se descurcă cel mai bine la Programare, care rămân fiabile și unde apar cele mai mari diferențe. Sortează după: Timp de răspuns (mediu) ↓.
| Rang | Model | Companie | Scor Programare | Scor | Teste corecte | Timp de răspuns (mediu) |
|---|---|---|---|---|---|---|
| #50 | GLM 5 none | Z.ai | 5.6 | 6.6 | 0/1 | 8.84s |
| #41 | GPT-5.4 Mini medium | OpenAI | 10.0 | 7.3 | 1/1 | 7.76s |
| #61 | DeepSeek V3.2 none | DeepSeek | 2.4 | 6.1 | 0/1 | 7.63s |
| #57 | Gemma 4 26B A4B none | 4.7 | 6.2 | 0/1 | 7.07s | |
| #3 | Gemini 3 Flash Preview low | 10.0 | 8.8 | 1/1 | 6.94s | |
| #84 | Qwen3.5-9B none | Qwen | 5.2 | 4.8 | 0/1 | 5.69s |
| #70 | GLM 4.7 Flash none | Z.ai | 6.4 | 5.6 | 0/1 | 5.57s |
| #55 | GLM 5V Turbo none | Z.ai | 10.0 | 6.2 | 1/1 | 5.30s |
| #58 | Seed-2.0-Lite none | Bytedance Seed | 10.0 | 6.2 | 1/1 | 4.61s |
| #15 | Gemini 3.1 Flash Lite Preview medium | 10.0 | 8.2 | 1/1 | 4.34s | |
| #73 | GLM 5 Turbo none | Z.ai | 5.3 | 5.5 | 0/1 | 3.93s |
| #62 | MiMo-V2-Pro none | Xiaomi | 10.0 | 6.0 | 1/1 | 3.82s |
| #39 | Claude Sonnet 4.6 none | Anthropic | 10.0 | 7.4 | 1/1 | 3.67s |
| #46 | Qwen3.5 Plus 2026-02-15 none | Qwen | 6.3 | 6.8 | 0/1 | 3.63s |
| #66 | Qwen3.5-122B-A10B none | Qwen | 4.3 | 5.7 | 0/1 | 3.44s |