Claude Opus 5 (Xhigh) vs GPT-6 Luna (High)
Which is better for the money: Claude Opus 5 (Xhigh) by Anthropic or GPT-6 Luna (High) by OpenAI? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-26.
Trade-off
Verdict
Depends on your budget
Claude Opus 5 (Xhigh) scores 16.8 points higher (49.7 vs 32.9) but costs 123× as much to run the benchmark ($5868 vs $47.9; flagship vs budget cost classes). They are not alternatives at one budget: pick Claude Opus 5 (Xhigh) when you need intelligence above 32.9, and GPT-6 Luna (High) when 32.9 is enough.
| Metric | Claude Opus 5 (Xhigh) | GPT-6 Luna (High) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 49.7 | 32.9 | −33.8% GPT-6 Luna (High) is 33.8% less intelligent |
| Total AA cost (lower is better) | $5,867.76 | $47.85 | −99.2% GPT-6 Luna (High) is 99.2% cheaper |
| End-to-end time (lower is better) | 40.34s | 21.44s | −46.9% GPT-6 Luna (High) is 46.9% faster |
| Price per 1M tokens (in / out) | $5.00 / $25.0 | $0.10 / $0.50 | |
| Input modalities | Text, Image | Text, Image |
GPT-6 Luna (High) vs Claude Opus 5 (Xhigh): 33.8% less intelligence, 99.2% less cost to run the AA benchmark, and 46.9% faster end-to-end.
Today's Pareto frontier
Against the frontier
- Claude Opus 5 (Xhigh) is off the frontier: at the same or lower cost, Claude Opus 5.5 (Xhigh, Default Fallback) scores 56.00 for $4057.
- GPT-6 Luna (High) is on today's frontier: nothing is both smarter and cheaper.
Claude Opus 5 (Xhigh)GPT-6 Luna (High)Today's Pareto frontierEvery other model
Matchup
Compare something else
Agents:
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/claude-opus-5-xhigh/gpt-6-luna-high · Markdown: /api/v1/compare/claude-opus-5-xhigh/gpt-6-luna-high?format=md · llms.txt · agent skill