token.app › Compare › MiniMax M3 vs Beam 1.0 501B A23B
MiniMax M3 vs Beam 1.0 501B A23B
Beam 1.0 501B A23B is 1.2× cheaper per token on a blended 3:1 basis.
Head to head
Blended price uses a 3:1 input:output mix
| MiniMax M3 | Beam 1.0 501B A23B | |
|---|---|---|
| Provider | MiniMax | Reflection |
| Input $/1M | $0.300 | $0.250 |
| Output $/1M | $1.20 | $1.00 |
| Blended $/1M | $0.525 | $0.438 |
| Context window | 1.0M | 1.0M |
| Max output | 512K | 131K |
| Released | May 2026 | Oct 2026 |
| Licence | Open weights | Open weights |
Choose MiniMax M3 if…
- No clear advantage on the data we hold.
Choose Beam 1.0 501B A23B if…
- Costs less per token — $0.438 vs $0.525 blended
- Newer model (released Oct 2026)
FAQ
Which is better, MiniMax M3 or Beam 1.0 501B A23B?
We do not hold independently-run benchmark scores covering both MiniMax M3 and Beam 1.0 501B A23B, so we make no quality claim. Their pricing and specifications are compared on this page.
Is MiniMax M3 cheaper than Beam 1.0 501B A23B?
Beam 1.0 501B A23B is cheaper: $0.438 versus $0.525 per million tokens blended at 3:1 input:output. Input alone: $0.300 vs $0.250. Output alone: $1.20 vs $1.00.
What context windows do MiniMax M3 and Beam 1.0 501B A23B have?
MiniMax M3: 1,048,576 tokens. Beam 1.0 501B A23B: 1,048,576 tokens.
Who makes MiniMax M3 and Beam 1.0 501B A23B?
MiniMax M3 is made by MiniMax. Beam 1.0 501B A23B is made by Reflection.