7.4 Global

qwen2.5:7b

Judge: gemma4:31b · 144/160 tests · 16 min 5 s · 44.0 tok/s

7.6B · Q4_K_M · 4.4 GB · 33K ctx

Tools

Desglose por categoría

surprise 10.0

agentic 9.0

instruction 8.6

long-context 8.4

multilingual 8.0

code 7.5

reasoning 7.3

math 7.2

roleplay 7.1

frontend 6.9

safety 6.8

organization 6.0

writing 5.7