GPT-5.6 Luna (High) vs Beam 1.0
Which is better for the money: GPT-5.6 Luna (High) by OpenAI or Beam 1.0 by Reflection? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-24.
Clear winner
Verdict
Pick GPT-5.6 Luna (High)
Beam 1.0 vs GPT-5.6 Luna (High): 5.6% less intelligence, 194.7% more cost to run the AA benchmark, and 26.4% faster end-to-end.
| Metric | GPT-5.6 Luna (High) | Beam 1.0 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 32.1 | 30.3 | −5.6% Beam 1.0 is 5.6% less intelligent |
| Total AA cost (lower is better) | $108.15 | $318.73 | +194.7% Beam 1.0 is 2.9× more expensive |
| End-to-end time (lower is better) | 10.89s | 8.02s | −26.4% Beam 1.0 is 26.4% faster |
| Price per 1M tokens (in / out) | $0.20 / $1.20 | $0.40 / $1.20 | |
| Input modalities | Text, Image | Text |
Today's Pareto frontier
Against the frontier
- GPT-5.6 Luna (High) is off the frontier: at the same or lower cost, Claude Haiku 5.5 (High) scores 37.80 for $94.1.
- Beam 1.0 is off the frontier: at the same or lower cost, MiMo-V2.6-Pro scores 46.30 for $207.
GPT-5.6 Luna (High)Beam 1.0Today'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-high/reflection-beam · Markdown: /api/v1/compare/gpt-5-6-luna-high/reflection-beam?format=md · llms.txt · agent skill