List your company

Multiple-choice Question Answering on HLE filtered subset

19.6NCommGroupDebate
-0.7844.5089.815.092May 21, 2026

Evaluation results

MethodLinks
GroupDebateLLM=DeepSeek-V3.12026.0519.6108.714.9
S^2-MADLLM=DeepSeek-V3.12026.0518.771.716.7
GroupDebateLLM=GPT-OSS-120B2026.0518.411120.2
SID-ETLLM=DeepSeek-V3.12026.0516.633.714.1
S^2-MADLLM=GPT-OSS-120B2026.051467.721
SID-ETLLM=GPT-OSS-120B2026.0512.839.617.2
SVR-MADLLM=GPT-OSS-120B2026.057.339.121.4
SVR-MADLLM=DeepSeek-V3.12026.055.330.316.7
Self ConsistencyLLM=GPT-OSS-120B2026.05014.113.9
Self ConsistencyLLM=DeepSeek-V3.12026.0509.814.1