modelgrep

Qwen: Qwen3 Next 80B A3B Thinking vs Z.ai: GLM 4.5

Qwen: Qwen3 Next 80B A3B Thinking wins on more metrics (6 of 9), but the right pick depends on what you optimize for — see the breakdown below.

MetricQwen: Qwen3 Next 80B A3B ThinkingZ.ai: GLM 4.5
Intelligence Index26.726.4
Coding Index19.526.3
GPQA Diamond76%78%
Design Arena Elo1251
Speed (tokens/sec)17240
Latency252ms2.4s
Input price /M$0.098$0.600
Output price /M$0.780$2.20
Context window262K131K
CapabilitiesReasoningToolsJSONReasoningToolsJSON