Share your thoughts, 1 month free Claude Pro on usSee more

Agentic Reasoning on PinchBench

91.3Score

DS-v4-Flash

Updated 1mo ago

Evaluation Results

Method	Links
DS-v4-Flash 2026.06		91.3
Kimi-K2.6 2026.06		90.2
Nemotron 3 Ultra 2026.06		90
DS-v4-Pro 2026.06		88.6
Qwen-3.5 2026.06		86.6
Kimi-K2.5 2026.06		85.48
DeepSeek-V3.2 2026.06		85.38
Ling-2.6-1T 2026.06		85.24
GLM-5 2026.06		83.29
Ling-2.6-flash 2026.06		81.3
GLM-5.1 2026.06		81.2
MiniMax-2.7 2026.06		77.6
GPT-5.4 2026.06		73.4
Nemotron-3-Super 120B-A12B 2026.06		73.1
GPT-5.4-mini 2026.06		71.4
GPT-OSS-120B 2026.06		52