List your company

Same-task discrimination accuracy on BBEH

72.6Same-task AccuracyOE
47.95254.35160.7567.149May 7, 2026

Evaluation results

MethodLinks
OELLM=Llama 70B2026.0572.6
OELLM=Llama 8B2026.0568.8
OELLM=DS-R1 32B2026.0563.5
OELLM=Sky-OR1 32B2026.0563.4
OELLM=GPT-OSS 20B2026.0562.4
STLLM=DS-R1 32B2026.0554.8
STLLM=Llama 8B2026.0554.7
LTLLM=Llama 8B2026.0554.7
IELLM=Llama 8B2026.0554.5
STLLM=Sky-OR1 32B2026.0554.2
IELLM=Sky-OR1 32B2026.0554
LITLLM=Llama 8B2026.0553.9
STNLLM=Llama 8B2026.0553.5
IELLM=DS-R1 32B2026.0553.5
LITLLM=Sky-OR1 32B2026.0553.3
STLLM=Llama 70B2026.0552.9
STNLLM=Llama 70B2026.0552.8
LITLLM=DS-R1 32B2026.0552.6
IELLM=Llama 70B2026.0552.1
STNLLM=DS-R1 32B2026.0551.8
STNLLM=Sky-OR1 32B2026.0551.8
STNLLM=GPT-OSS 20B2026.0551.8
LITLLM=Llama 70B2026.0551.6
LTLLM=Llama 70B2026.0551.4
LTLLM=Sky-OR1 32B2026.0551.2
LTLLM=GPT-OSS 20B2026.0551.2
RVLLM=GPT-OSS 20B2026.0551
STLLM=GPT-OSS 20B2026.0550.7
LITLLM=GPT-OSS 20B2026.0550.4
LTLLM=DS-R1 32B2026.0550.3
RVLLM=DS-R1 32B2026.0550.3
RVLLM=Sky-OR1 32B2026.0549.9
RVLLM=Llama 70B2026.0549.7
IELLM=GPT-OSS 20B2026.0549.5
RVLLM=Llama 8B2026.0548.9