Compare Models
Select models (max 7)
Gemini 4 CarbonGPT-6.1 Terra
 |  |  |
|---|---|---|
| Vals IndexGDP-weighted benchmark | 70.36%±0.97 | 60.47%±0.99 |
| Legal Research BenchAgentic US legal research | 55.77%±3.45 | 25.96%±3.05 |
| Finance Agent v2.1Core financial analyst tasks | 67.16%±2.02 | 55.96%±2.09 |
| Tax Agent BenchAgentic US corporate tax research | 48.86%±3.42 | 15.17%±2.43 |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | 63.13%±2.20 | 49.49%±1.82 |
| Vibe Code Bench v1.1Building web apps from scratch | 89.06%±2.29 | 84.00%±2.87 |
Benchmarks
Vals Index *
Gemini 4 Carbon
0.00%± 0.97
(29/29)GPT-6.1 Terra
0.00%± 0.99
(29/29)Legal Research Bench *
Gemini 4 Carbon
0.00%± 3.45
(87/87)GPT-6.1 Terra
0.00%± 3.05
(87/87)Finance Agent v2.1 *
Gemini 4 Carbon
0.00%± 2.02
(33/33)GPT-6.1 Terra
0.00%± 2.09
(33/33)Tax Agent Bench *
Gemini 4 Carbon
0.00%± 3.42
(78/78)GPT-6.1 Terra
0.00%± 2.43
(78/78)Terminal-Bench 4.0
Gemini 4 Carbon
0.00%± 2.20
(57/57)GPT-6.1 Terra
0.00%± 1.82
(57/57)Vibe Code Bench v1.1 *
Gemini 4 Carbon
0.00%± 2.29
(121/121)GPT-6.1 Terra
0.00%± 2.87
(121/121)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, and law
•2026-10-27
Comparison by Industry
Model performance on different sections of the economy.
Legal
36.63%
vs.13.81%
Finance
64.16%
vs.48.15%
Coding
76.10%
vs.66.75%
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
$10.51
$2.925
Cost / TestVals Index
$10.51
vs.$2.925
Input Cost/ 1M Tokens
$4.00
vs.$1.00
Input Cache Write/ 1M Tokens
N/A
vs.$1.25
Input Cache Read/ 1M Tokens
$0.20
vs.$0.05
Output Cost/ 1M Tokens
$20.00
vs.$5.00
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | Google | OpenAI |
| Latency | 2864.86s | 2374.82s |
| Cost (In/Out) | $4 / $20 | $1 / $5 |
| Context Window | 1M | 1M |
| Max Output Token | 1M | 128k |
| Input Modality |