LiveCodeBench - August 2024 to February 2025

Code

LiveCodeBench code-generation problems in the August 1, 2024 to February 1, 2025 date window, labelled "24.08-25.02" by Qwen, "2024-08 to 2025-01" in the DeepSeek-R1 report and 20240801-20250201 or 2408-2502 by NVIDIA. This exact historical problem cohort is retained as a diagnostic rather than pooled with other LiveCodeBench windows.

Top models (higher is better)

Model	Score
Nemotron Cascade 14B Thinking	77.5
Nemotron Cascade 8B Thinking	74.5
Nemotron Cascade 8B	74.3
Nemotron Nano 9B V2	68.2
DeepSeek-R1	65.9
Qwen3 14B	65.2
QwQ-32B	63.4
Qwen3 8B	61.2
DeepSeek-R1-Distill-Llama-70B	57.5
DeepSeek-R1-Distill-Qwen-32B	57.2
o1-mini	53.8
DeepSeek R1 Distill Qwen 14B	53.1
DeepSeek-R1-Zero	50.0
Llama 3.1 Nemotron Nano 8B V1	46.6
Llama 3.3 Nemotron Super 49B V1	45.5
QwQ-32B-Preview	41.9
DeepSeek R1 Distill Llama 8B	39.6
DeepSeek-R1-Distill-Qwen-7B	37.6
DeepSeek-V3	36.2
DeepSeek-R1-Distill-Qwen-1.5B	16.9
Llama 3.1 8B Instruct	11.8