modelgrep

Anthropic: Claude Sonnet 4 vs DeepSeek: DeepSeek V3.2 Exp

DeepSeek: DeepSeek V3.2 Exp wins on more metrics (5 of 9), but the right pick depends on what you optimize for — see the breakdown below.

MetricAnthropic: Claude Sonnet 4DeepSeek: DeepSeek V3.2 Exp
Intelligence Index33.032.1
Coding Index30.634.6
GPQA Diamond68%75%
Design Arena Elo12201229
Speed (tokens/sec)5020
Latency720ms1.3s
Input price /M$3.00$0.270
Output price /M$15.00$0.410
Context window1M164K
CapabilitiesReasoningToolsVisionReasoningToolsJSON