Data Prediction on MLEBench lite benchmark leaderboard · SOTA2 Research

Data Prediction on MLEBench lite

100Valid ScoreQwen3-Coder

Evaluation Results

Method	Links
Qwen3-Coder2026.01		100	9.09	22.72
GPT-5.12026.01		90.91	22.73	45.45
GPT-5.12026.01		90.91	22.73	31.82
Claude Sonnet 42026.01		90.9	13.63	22.73
Kimi K2 Instruct2026.01		86.37	13.64	27.27
Deepseek v3.12026.01		86.37	13.64	27.27
Claude Sonnet 4.52026.01		86.36	22.73	36.36
Qwen3 Instruct2026.01		81.82	4.55	13.64
GPT-52026.01		77.27	9.09	27.27
GPT-5.12026.01		72.72	13.64	22.73
Qwen3 Instruct2026.01		50	4.55	9.09
Qwen2.5 Instruct2026.01		0	0	0