Beam 1.0 501 B A23 B vs Moondream 3 Preview: cost comparison
Prices updated 2026-11-06 · computed from live provider data
Short answer: Beam 1.0 501 B A23 B is cheaper on a
3:1 input/output blend — $0.475 per million tokens versus
$0.850, a 44% difference.
Beam 1.0 501 B A23 B offers the larger context window at 1,048,576 tokens, against 32,000 for Moondream 3 Preview.
Head to head
| Beam 1.0 501 B A23 B | Moondream 3 Preview | |
|---|---|---|
| Input, per 1M tokens | $0.300 | $0.300 |
| Output, per 1M tokens | $1.00 | $2.50 |
| Blended 3:1, per 1M | $0.475 | $0.850 |
| Context window | 1,048,576 | 32,000 |
| Vendor | reflection | moondream |
Monthly cost at real volumes
| Workload (input / output) | Beam 1.0 501 B A23 B | Moondream 3 Preview |
|---|---|---|
| 1M in / 250K out | $0.550 | $0.925 |
| 10M in / 2.5M out | $5.50 | $9.25 |
| 100M in / 25M out | $55.00 | $92.50 |
Which should you pick?
Pick Beam 1.0 501 B A23 B when you need the longer context window (1,048,576 vs 32,000 tokens). Pick Beam 1.0 501 B A23 B when price per token dominates your workload — high-volume classification, extraction, or batch summarisation.