Compare Models
Select models (max 5)
Beam 1.0MiniMax-M3.1-Flash
 |  |  |
|---|---|---|
| Vals IndexGDP-weighted benchmark | 43.42%±1.06 | 53.78%±1.07 |
| Legal Research BenchAgentic US legal research | 28.85%±3.15 | 34.13%±3.30 |
| Finance Agent v2.1Core financial analyst tasks | 51.83%±2.11 | 58.83%±2.07 |
| Tax Agent BenchAgentic US corporate tax research | 8.81%±1.90 | 32.14%±3.34 |
| Code Migration v1.1Rewriting programs in new languages | 22.85%±3.28 | N/A |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | 4.04%±1.34 | 25.25%±1.01 |
| Vibe Code Bench v1.1Building web apps from scratch | 66.44%±4.06 | 73.94%±4.25 |
Benchmarks
Vals Index *
Beam 1.0
0.00%± 1.06
(23/23)MiniMax-M3.1-Flash
0.00%± 1.07
(23/23)Legal Research Bench *
Beam 1.0
0.00%± 3.15
(82/82)MiniMax-M3.1-Flash
0.00%± 3.30
(82/82)Finance Agent v2.1 *
Beam 1.0
0.00%± 2.11
(28/28)MiniMax-M3.1-Flash
0.00%± 2.07
(28/28)Tax Agent Bench *
Beam 1.0
0.00%± 1.90
(73/73)MiniMax-M3.1-Flash
0.00%± 3.34
(73/73)Code Migration v1.1 *
Beam 1.0
0.00%± 3.28
(26/26)MiniMax-M3.1-Flash
N/ATerminal-Bench 4.0
Beam 1.0
0.00%± 1.34
(52/52)MiniMax-M3.1-Flash
0.00%± 1.01
(52/52)Vibe Code Bench v1.1 *
Beam 1.0
0.00%± 4.06
(116/116)MiniMax-M3.1-Flash
0.00%± 4.25
(116/116)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, and law
•2026-10-19
Comparison by Industry
Model performance on different sections of the economy.
Legal
16.71%
vs.20.82%
Finance
41.79%
vs.52.94%
Coding
31.11%
vs.49.60%
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
N/A
$0.66
Cost / TestVals Index
N/A
vs.$0.66
Input Cost/ 1M Tokens
N/A
vs.$0.15
Input Cache Read/ 1M Tokens
N/A
vs.$0.015
Output Cost/ 1M Tokens
N/A
vs.$0.60
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | Reflection | MiniMax |
| Latency | 446.56s | 2597.84s |
| Cost (In/Out) | N/A | $0.15 / $0.6 |
| Context Window | 1M | 1M |
| Max Output Token | 131k | 524k |
| Input Modality |