GPT-5.6 Luna (Non-reasoning) vs GPT-6 Astra (Medium)
Which is better for the money: GPT-5.6 Luna (Non-reasoning) by OpenAI or GPT-6 Astra (Medium) 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-6 Astra (Medium) scores 34.1 points higher (49.6 vs 15.5) but costs 106× as much to run the benchmark ($2434 vs $23.0; flagship vs budget cost classes). They are not alternatives at one budget: pick GPT-6 Astra (Medium) when you need intelligence above 15.5, and GPT-5.6 Luna (Non-reasoning) when 15.5 is enough.
| Metric | GPT-5.6 Luna (Non-reasoning) | GPT-6 Astra (Medium) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 15.5 | 49.6 | +220.0% GPT-6 Astra (Medium) is 3.2× more intelligent |
| Total AA cost (lower is better) | $22.97 | $2,434.12 | +10497.0% GPT-6 Astra (Medium) is 106× more expensive |
| End-to-end time (lower is better) | 5.06s | 14.89s | +194.3% GPT-6 Astra (Medium) is 194.3% slower |
| Price per 1M tokens (in / out) | $0.20 / $1.20 | $10.0 / $50.0 | |
| Input modalities | Text, Image | Text, Image |
GPT-6 Astra (Medium) vs GPT-5.6 Luna (Non-reasoning): 220.0% more intelligence, 10497.0% more cost to run the AA benchmark, and 194.3% slower end-to-end.
Today's Pareto frontier
Against the frontier
- GPT-5.6 Luna (Non-reasoning) is off the frontier: at the same or lower cost, GPT-6 Luna (Low) scores 21.50 for $10.6.
- GPT-6 Astra (Medium) is off the frontier: at the same or lower cost, Claude Opus 5.5 (High, Default Fallback) scores 53.60 for $2172.
GPT-5.6 Luna (Non-reasoning)GPT-6 Astra (Medium)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/gpt-5-6-luna-non-reasoning/gpt-6-astra-medium · Markdown: /api/v1/compare/gpt-5-6-luna-non-reasoning/gpt-6-astra-medium?format=md · llms.txt · agent skill