modelgrep

Anthropic: Claude Opus 4.8 vs OpenAI: GPT-5.2-Codex

anthropic logoAnthropic: Claude Opus 4.8anthropic/claude-opus-4.8VSopenai logoOpenAI: GPT-5.2-Codexopenai/gpt-5.2-codex

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-5.2-Codex
Intelligence Index55.7✓40.1
Coding Index74.3✓—
GPQA Diamond92%✓90%
Design Arena Elo1288✓—
Speed (tokens/sec)72✓64
Latency1.8s1.6s✓
Input price /M$5.00$1.75✓
Output price /M$25.00$14.00✓
Context window1M✓400K
CapabilitiesReasoningToolsJSONVisionReasoningToolsJSONVision
Anthropic: Claude Opus 4.8 details →OpenAI: GPT-5.2-Codex details →