DeepSeek V3 (Dec '24) vs GPT-5.6 Luna (Xhigh)
Which is better for the money: DeepSeek V3 (Dec '24) by DeepSeek or GPT-5.6 Luna (Xhigh) 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
GPT-5.6 Luna (Xhigh) scores 26.1 points higher (34.6 vs 8.5) but costs 7.1× as much to run the benchmark ($179 vs $25.4; mid-range vs budget cost classes). They are not alternatives at one budget: pick GPT-5.6 Luna (Xhigh) when you need intelligence above 8.5, and DeepSeek V3 (Dec '24) when 8.5 is enough.
| Metric | DeepSeek V3 (Dec '24) | GPT-5.6 Luna (Xhigh) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 8.5 | 34.6 | +307.1% GPT-5.6 Luna (Xhigh) is 4.1× more intelligent |
| Total AA cost (lower is better) | $25.37 | $179.33 | +606.9% GPT-5.6 Luna (Xhigh) is 7.1× more expensive |
| End-to-end time (lower is better) | 12.35s | 35.85s | +190.3% GPT-5.6 Luna (Xhigh) is 190.3% slower |
| Price per 1M tokens (in / out) | $0.32 / $0.89 | $0.20 / $1.20 | |
| Input modalities | Text | Text, Image |
GPT-5.6 Luna (Xhigh) vs DeepSeek V3 (Dec '24): 307.1% more intelligence, 606.9% more cost to run the AA benchmark, and 190.3% slower end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek V3 (Dec '24) is off the frontier: at the same or lower cost, GPT-6 Luna (Low) scores 21.50 for $10.6.
- GPT-5.6 Luna (Xhigh) is off the frontier: at the same or lower cost, Claude Haiku 5.5 (Xhigh) scores 41.20 for $158.
DeepSeek V3 (Dec '24)GPT-5.6 Luna (Xhigh)Today's Pareto frontierOpen-weights frontierEvery other model
Matchup
Compare something else
Agents:
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/deepseek-v3/gpt-5-6-luna-xhigh · Markdown: /api/v1/compare/deepseek-v3/gpt-5-6-luna-xhigh?format=md · llms.txt · agent skill