Coding
LiveCodeBench
Problemas de coding contests en vivo de LeetCode/Codeforces.
30 modelos publicaron score
| # | Modelo | Empresa | Score |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | DeepSeek | 93.5 |
| 2 | Kimi K2.6 | Moonshot AI | 89.6 |
| 3 | DeepSeek V3.2 Speciale | DeepSeek | 88.7 |
| 4 | Doubao Seed 2.0 Pro | ByteDance | 87.8 |
| 5 | Step 3.5 Flash | StepFun | 86.4 |
| 6 | Qwen3-Max-Thinking | Alibaba | 85.9 |
| 7 | Kimi K2.5 | Moonshot AI | 85.0 |
| 8 | Qwen3.6-27B | Alibaba | 83.9 |
| 9 | Qwen3.5-397B-A17B | Alibaba | 83.6 |
| 10 | DeepSeek V3.2 | DeepSeek | 83.3 |
| 11 | EXAONE 4.5 33B | LG AI Research | 81.4 |
| 12 | Nemotron 3 Super | Nvidia | 81.2 |
| 13 | K-EXAONE 236B-A23B | LG AI Research | 80.7 |
| 14 | Qwen3.6-35B-A3B | Alibaba | 80.4 |
| 15 | Gemma 4 (31B dense) | Google DeepMind | 80.0 |
| 16 | Grok 4 | xAI | 79.4 |
| 17 | Gemma 4 26B-A4B | Google DeepMind | 77.1 |
| 18 | Magistral Medium 1.2 | Mistral AI | 75.0 |
| 19 | DeepSeek R1 0528 | DeepSeek | 73.3 |
| 20 | GLM-4.6 | Zhipu AI | 70.1 |
| 21 | Nemotron 3 Nano | Nvidia | 68.3 |
| 22 | Step-3 | StepFun | 67.1 |
| 23 | MiniMax M2.5 | MiniMax | 65.0 |
| 24 | Reka Flash 3.1 | Reka | 53.5 |
| 25 | Gemma 4 E4B | Google DeepMind | 52.0 |
| 26 | Gemma 4 E2B | Google DeepMind | 44.0 |
| 27 | Reka Flash 3 | Reka | 43.5 |
| 28 | Llama 4 Maverick | Meta | 43.4 |
| 29 | Codestral 25.08 | Mistral AI | 37.9 |
| 30 | Llama 4 Scout | Meta | 32.8 |