Compare Models
Select models (max 5)
Gemini 4 ArgonGPT-6 Astra
 |  |  |
|---|---|---|
| Vals IndexGDP-weighted benchmark | 68.90%±0.97 | 63.13%±1.19 |
| Legal Research BenchAgentic US legal research | 54.81%±3.46 | 39.42%±3.40 |
| Finance Agent (v2)Core financial analyst tasks | 65.40%±0.32 | 53.54%±2.08 |
| Tax Agent BenchAgentic US corporate tax research | 44.56%±3.59 | 20.72%±2.92 |
| MedCodeMedical billing code support | 58.80%±2.10 | 48.49%±2.13 |
| Terminal-Bench-ScienceExpert-authored scientific research workflows | 44.29%±5.98 | 62.86%±5.82 |
| Code MigrationRewriting programs in new languages | 68.17%±4.35 | 67.74%±4.22 |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | 57.58%±2.31 | 59.60%±4.40 |
| Vibe Code Bench v1.1Building web apps from scratch | 91.91%±1.90 | 89.59%±2.17 |
Benchmarks
Vals Index *
Gemini 4 Argon
0.00%± 0.97
(50/50)GPT-6 Astra
0.00%± 1.19
(50/50)Legal Research Bench *
Gemini 4 Argon
0.00%± 3.46
(80/80)GPT-6 Astra
0.00%± 3.40
(80/80)Finance Agent (v2) *
Gemini 4 Argon
0.00%± 0.32
(80/80)GPT-6 Astra
0.00%± 2.08
(80/80)Tax Agent Bench *
Gemini 4 Argon
0.00%± 3.59
(71/71)GPT-6 Astra
0.00%± 2.92
(71/71)MedCode *
Gemini 4 Argon
0.00%± 2.10
(107/107)GPT-6 Astra
0.00%± 2.13
(107/107)Terminal-Bench-Science
Gemini 4 Argon
0.00%± 5.98
(41/41)GPT-6 Astra
0.00%± 5.82
(41/41)Code Migration *
Gemini 4 Argon
0.00%± 4.35
(79/79)GPT-6 Astra
0.00%± 4.22
(79/79)Terminal-Bench 4.0
Gemini 4 Argon
0.00%± 2.31
(50/50)GPT-6 Astra
0.00%± 4.40
(50/50)Vibe Code Bench v1.1 *
Gemini 4 Argon
0.00%± 1.90
(114/114)GPT-6 Astra
0.00%± 2.17
(114/114)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, and law
•2026-10-18
Comparison by Industry
Model performance on different sections of the economy.
Legal
54.23%
vs.22.42%
Finance
61.73%
vs.48.65%
Healthcare
73.11%
vs.68.20%
Math
99.00%
vs.99.00%
Science
41.73%
vs.54.65%
Education
53.65%
vs.46.37%
Coding
56.77%
vs.56.20%
Cyber
61.07%
vs.48.97%
Social Mobility
69.76%
vs.N/A
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
$15.68
$18.46
Cost / TestVals Index
$15.68
vs.$18.46
Input Cost/ 1M Tokens
$4.00
vs.$10.00
Input Cache Write/ 1M Tokens
N/A
vs.$12.50
Input Cache Read/ 1M Tokens
$0.20
vs.$1.00
Output Cost/ 1M Tokens
$20.00
vs.$50.00
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | Google | OpenAI |
| Latency | 2792.77s | 1667.74s |
| Cost (In/Out) | $4 / $20 | $10 / $50 |
| Context Window | 1M | 1M |
| Max Output Token | 262k | 128k |
| Input Modality |