modelgrep

Anthropic: Claude Opus 4.6 vs Anthropic: Claude Opus 4.7

Anthropic: Claude Opus 4.7 wins on more metrics (5 of 6), but the right pick depends on what you optimize for — see the breakdown below.

MetricAnthropic: Claude Opus 4.6Anthropic: Claude Opus 4.7
Intelligence Index52.957.3
Coding Index48.152.5
GPQA Diamond90%91%
Design Arena Elo13581503
Speed (tokens/sec)4163
Latency1.5s1.6s
Input price /M$5.00$5.00
Output price /M$25.00$25.00
Context window1M1M
CapabilitiesReasoningToolsJSONVisionReasoningToolsJSONVision