modelbenchmark.io

openPangu 2.0 Flash

openpangu-2-0-flash

Compare

Huawei's 92B MoE lightweight (6B active) for fast agents and 512K context work, Ascend-native with slow/fast thinking

Specification

most-agreed values

Context
524K
Max output
131K
Released
2026-06-30
Knowledge cutoff
2026-06
Retires
—
Open weights
yes
Input
text
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.1199
Output
$0.2399
Cache read
$0.03

Available from 1 host

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
huawei-maas$0.1199$0.2399$0.03—524K131K—