LiveCodeBench - July to November 2024

Code

LiveCodeBench code-generation problems in the July 2024 to November 2024 date window that Qwen labels "LiveCodeBench 2024 07 - 2024 11". This exact historical problem cohort is retained as a diagnostic rather than pooled with other LiveCodeBench windows.

Top models (higher is better)

Model	Score
GPT-4o (2024-08-06)	34.6
Claude 3.5 Sonnet (Oct 2024)	31.6
Qwen2.5 Coder 32B Instruct	31.4
Qwen2.5 Coder 14B Instruct	23.4
Codestral 2405	22.6
Qwen2.5-Coder-7B-Instruct	18.2
Qwen2.5 Coder 3B Instruct	10.8
Qwen2.5 Coder 1.5B Instruct	6.1
Qwen2.5 Coder 0.5B Instruct	2.0