modelgrep

Qwen: Qwen3 235B A22B Thinking 2507 vs Z.ai: GLM 4.6

Qwen: Qwen3 235B A22B Thinking 2507 wins on more metrics (6 of 9), but the right pick depends on what you optimize for — see the breakdown below.

MetricQwen: Qwen3 235B A22B Thinking 2507Z.ai: GLM 4.6
Intelligence Index29.530.2
Coding Index23.230.2
GPQA Diamond79%63%
Design Arena Elo10971219
Speed (tokens/sec)7239
Latency508ms658ms
Input price /M$0.100$0.430
Output price /M$0.100$1.74
Context window262K203K
CapabilitiesReasoningToolsJSONReasoningToolsJSON