modelgrep

DeepSeek: DeepSeek V4 Flash vs Qwen: Qwen3.7 Max

DeepSeek: DeepSeek V4 Flash wins on more metrics (5 of 9), but the right pick depends on what you optimize for — see the breakdown below.

MetricDeepSeek: DeepSeek V4 FlashQwen: Qwen3.7 Max
Intelligence Index40.346.0
Coding Index38.750.1
GPQA Diamond89%92%
Design Arena Elo12781334
Speed (tokens/sec)6844
Latency550ms1.7s
Input price /M$0.098$1.25
Output price /M$0.196$3.75
Context window1.0M1M
CapabilitiesReasoningToolsJSONReasoningToolsJSON