List your company
ResearchFormerly WizWand

Sub-concept Matching on SCAR

78System AccuracyGrok-4
35.3646.4357.568.57May 22, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Grok-4Background Descriptions=-BG2026.0578
Grok-4Background Descriptions=+BG2026.0576
DeepSeek-R1Background Descriptions=-BG2026.0576
GPT-OSS-120BBackground Descriptions=-BG2026.0572
Qwen3-32BBackground Descriptions=-BG2026.0571
DeepSeek-R1Background Descriptions=+BG2026.0570
Qwen3-14BBackground Descriptions=-BG2026.0570
GPT-OSS-120BBackground Descriptions=+BG2026.0569
Qwen3-14BBackground Descriptions=+BG2026.0568
GPT-OSS-20BBackground Descriptions=-BG2026.0568
GPT-4.1-miniBackground Descriptions=-BG2026.0568
GPT-OSS-20BBackground Descriptions=+BG2026.0566
GPT-4.1-miniBackground Descriptions=+BG2026.0566
Qwen3-32BBackground Descriptions=+BG2026.0565
Gemini-2.5Background Descriptions=-BG2026.0564
Llama-405BBackground Descriptions=-BG2026.0562
Llama-70BBackground Descriptions=-BG2026.0560
Llama-70BBackground Descriptions=+BG2026.0557
Llama-405BBackground Descriptions=+BG2026.0556
Gemini-2.5Background Descriptions=+BG2026.0554
Llama-8BBackground Descriptions=-BG2026.0546
GPT-4.1-nanoBackground Descriptions=-BG2026.0546
Llama-8BBackground Descriptions=+BG2026.0540
GPT-4.1-nanoBackground Descriptions=+BG2026.0537