Muse Ember (Xhigh) vs Beam 1.0
Which is better for the money: Muse Ember (Xhigh) by Meta or Beam 1.0 by Reflection? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-11-09.
Clear winner
Verdict
Pick Muse Ember (Xhigh)
Beam 1.0 vs Muse Ember (Xhigh): 48.1% less intelligence, 161.0% more cost to run the AA benchmark, and end-to-end time not fully reported.
| Metric | Muse Ember (Xhigh) | Beam 1.0 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 34.1 | 17.7 | −48.1% Beam 1.0 is 48.1% less intelligent |
| Total AA cost (lower is better) | $95.73 | $249.83 | +161.0% Beam 1.0 is 2.6× more expensive |
| End-to-end time (lower is better) | Not reported | 7.01s | — not comparable |
| Price per 1M tokens (in / out) | $0.10 / $0.50 | $0.40 / $1.20 | |
| Input modalities | Text, Image | Text |
Today's Pareto frontier
Against the frontier
- Muse Ember (Xhigh) is on today's frontier: nothing is both smarter and cheaper.
- Beam 1.0 is off the frontier: at the same or lower cost, GPT-6.1 Terra (Medium) scores 38.20 for $190.
Muse Ember (Xhigh)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/muse-ember-xhigh/reflection-beam · Markdown: /api/v1/compare/muse-ember-xhigh/reflection-beam?format=md · llms.txt · agent skill