GPT-5.6 Luna (Non-reasoning) vs Qwen3.8 2.4T A95B
Which is better for the money: GPT-5.6 Luna (Non-reasoning) by OpenAI or Qwen3.8 2.4T A95B by Alibaba? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-27.
Trade-off
Verdict
Depends on your budget
Qwen3.8 2.4T A95B scores 24.4 points higher (39.9 vs 15.5) but costs 115× as much to run the benchmark ($2632 vs $23.0; flagship vs budget cost classes). They are not alternatives at one budget: pick Qwen3.8 2.4T A95B 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) | Qwen3.8 2.4T A95B | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 15.5 | 39.9 | +157.4% Qwen3.8 2.4T A95B is 2.6× more intelligent |
| Total AA cost (lower is better) | $22.97 | $2,632.05 | +11358.6% Qwen3.8 2.4T A95B is 115× more expensive |
| End-to-end time (lower is better) | 5.14s | 64.93s | +1163.2% Qwen3.8 2.4T A95B is 1163.2% slower |
| Price per 1M tokens (in / out) | $0.20 / $1.20 | $2.00 / $6.00 | |
| Input modalities | Text, Image | Text |
Qwen3.8 2.4T A95B vs GPT-5.6 Luna (Non-reasoning): 157.4% more intelligence, 11358.6% more cost to run the AA benchmark, and 1163.2% 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.
- Qwen3.8 2.4T A95B 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)Qwen3.8 2.4T A95BToday'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/gpt-5-6-luna-non-reasoning/qwen3-8-2-4t-a95b · Markdown: /api/v1/compare/gpt-5-6-luna-non-reasoning/qwen3-8-2-4t-a95b?format=md · llms.txt · agent skill