Compare Models
Select models (max 5)
Claude Sonnet 5.5GPT-6 Sol
 |  |  |
|---|---|---|
| Vals IndexGDP-weighted benchmark | 69.22%±0.96 | 62.57%±0.95 |
| Legal Research BenchAgentic US legal research | 48.08%±3.47 | 28.85%±3.15 |
| Finance Agent (v2)Core financial analyst tasks | 58.10%±0.67 | 49.05%±0.58 |
| Tax Agent BenchAgentic US corporate tax research | 73.39%±2.98 | 53.05%±3.29 |
| MedCodeMedical billing code support | 52.92%±2.12 | 47.07%±2.12 |
| Terminal-Bench ScienceExpert-authored scientific research workflows | 38.57%±5.86 | 31.43%±5.59 |
| Code MigrationRewriting programs in new languages | 69.83%±4.26 | 57.20%±4.21 |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | 53.03%±1.51 | 34.34%±2.02 |
| Vibe Code Bench v1.1Building web apps from scratch | 92.39%±1.26 | 87.82%±2.53 |
Benchmarks
Vals Index *
Claude Sonnet 5.5
0.00%± 0.96
(66/66)GPT-6 Sol
0.00%± 0.95
(66/66)Legal Research Bench *
Claude Sonnet 5.5
0.00%± 3.47
(69/69)GPT-6 Sol
0.00%± 3.15
(69/69)Finance Agent (v2) *
Claude Sonnet 5.5
0.00%± 0.67
(70/70)GPT-6 Sol
0.00%± 0.58
(70/70)Tax Agent Bench *
Claude Sonnet 5.5
0.00%± 2.98
(61/61)GPT-6 Sol
0.00%± 3.29
(61/61)MedCode *
Claude Sonnet 5.5
0.00%± 2.12
(102/102)GPT-6 Sol
0.00%± 2.12
(102/102)Terminal-Bench Science
Claude Sonnet 5.5
0.00%± 5.86
(33/33)GPT-6 Sol
0.00%± 5.59
(33/33)Code Migration *
Claude Sonnet 5.5
0.00%± 4.26
(69/69)GPT-6 Sol
0.00%± 4.21
(69/69)Terminal-Bench 4.0
Claude Sonnet 5.5
0.00%± 1.51
(37/37)GPT-6 Sol
0.00%± 2.02
(37/37)Vibe Code Bench v1.1 *
Claude Sonnet 5.5
0.00%± 1.26
(104/104)GPT-6 Sol
0.00%± 2.53
(104/104)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, and law
•2026-09-27
Comparison by Industry
Model performance on different sections of the economy.
Vals Index
69.22%
vs.45.34%
Legal
25.50%
vs.15.26%
Finance
69.07%
vs.57.88%
Healthcare
72.01%
vs.64.55%
Math
100.00%
vs.83.00%
Science
56.26%
vs.45.47%
Education
51.77%
vs.44.79%
Coding
74.58%
vs.52.79%
Beta
44.87%
vs.77.98%
Social Mobility
67.19%
vs.56.63%
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
$20.80
$7.564
Cost / TestVals Index
$20.80
vs.$7.564
Input Cost/ 1M Tokens
$2.00
vs.$2.00
Input Cache Write/ 1M Tokens
$2.50
vs.$2.50
Input Cache Read/ 1M Tokens
$0.20
vs.$0.20
Output Cost/ 1M Tokens
$10.00
vs.$10.00
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | Anthropic | OpenAI |
| Latency | 4207.12s | 1593.73s |
| Cost (In/Out) | $2 / $10 | $2 / $10 |
| Context Window | 1M | 1M |
| Max Output Token | 128k | 128k |
| Input Modality |