modelgrep

Anthropic: Claude Opus 4.8 vs OpenAI: GPT-5.6 Luna

anthropic logoAnthropic: Claude Opus 4.8anthropic/claude-opus-4.8VSopenai logoOpenAI: GPT-5.6 Lunaopenai/gpt-5.6-luna

Anthropic: Claude Opus 4.8 wins on more metrics (6 of 9), but the right pick depends on what you optimize for — see the breakdown below.

MetricAnthropic: Claude Opus 4.8OpenAI: GPT-5.6 Luna
Intelligence Index55.7✓51.2
Coding Index74.3✓71.4
GPQA Diamond92%✓91%
Design Arena Elo1288✓—
Speed (tokens/sec)106✓60
Latency1.1s✓1.7s
Input price /M$5.00$1.00✓
Output price /M$25.00$6.00✓
Context window1M1.1M✓
CapabilitiesReasoningToolsJSONVisionReasoningToolsJSONVision
Anthropic: Claude Opus 4.8 details →OpenAI: GPT-5.6 Luna details →