modelgrep

Anthropic: Claude Opus 4.5 vs StepFun: Step 3.7 Flash

Anthropic: Claude Opus 4.5 wins on more metrics (5 of 9), but the right pick depends on what you optimize for — see the breakdown below.

MetricAnthropic: Claude Opus 4.5StepFun: Step 3.7 Flash
Intelligence Index43.142.6
Coding Index42.937.1
GPQA Diamond81%81%
Design Arena Elo12971232
Speed (tokens/sec)5874
Latency772ms1.8s
Input price /M$5.00$0.200
Output price /M$25.00$1.15
Context window200K256K
CapabilitiesReasoningToolsJSONVisionReasoningToolsJSONVision