Compare Models
Select models (max 7)
Claude Opus 4.8GPT-5.6 Terra
| Legal Research BenchAgentic US legal research | 43.75%±3.45 | 41.35%±3.42 |
| Tax Agent BenchAgentic US corporate tax research | 30.00%±3.30 | 30.04%±3.31 |
| MedCodeMedical billing code support | 53.22%±2.17 | 43.41%±2.17 |
| Terminal-Bench-ScienceExpert-authored scientific research workflows | 4.29%±2.44 | 10.00%±3.61 |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | 23.23%±1.34 | 22.73%±2.31 |
| Vibe Code Bench v1.1Building web apps from scratch | 82.72%±3.08 | 74.59%±4.20 |
Benchmarks
Legal Research Bench *
Claude Opus 4.8
0.00%± 3.45
(98/98)GPT-5.6 Terra
0.00%± 3.42
(98/98)Tax Agent Bench *
Claude Opus 4.8
0.00%± 3.30
(89/89)GPT-5.6 Terra
0.00%± 3.31
(89/89)MedCode *
Claude Opus 4.8
0.00%± 2.17
(107/107)GPT-5.6 Terra
0.00%± 2.17
(107/107)Terminal-Bench-Science
Claude Opus 4.8
0.00%± 2.44
(47/47)GPT-5.6 Terra
0.00%± 3.61
(47/47)Terminal-Bench 4.0
Claude Opus 4.8
0.00%± 1.34
(68/68)GPT-5.6 Terra
0.00%± 2.31
(68/68)Vibe Code Bench v1.1 *
Claude Opus 4.8
0.00%± 3.08
(132/132)GPT-5.6 Terra
0.00%± 4.20
(132/132)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, legal, and tax
Comparison by Industry
Model performance on different sections of the economy.
Legal
63.66%
vs.63.23%
Finance
58.52%
vs.57.84%
Healthcare
69.49%
vs.63.14%
Math
N/A
vs.74.00%
Science
4.29%
vs.10.00%
Academic
89.53%
vs.88.01%
Education
N/A
vs.87.04%
Coding
57.10%
vs.60.81%
Cyber
N/A
vs.40.81%
Social Mobility
68.13%
vs.66.68%
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
N/A
N/A
Cost / TestVals Index
N/A
vs.N/A
Input Cost/ 1M Tokens
$5.00
vs.$2.00
Input Cache Write/ 1M Tokens
$6.25
vs.$2.50
Input Cache Read/ 1M Tokens
$0.50
vs.$0.20
Output Cost/ 1M Tokens
$25.00
vs.$12.00
Model Metadata
Basic information about each model.
| Model provider | ||
| Latency | N/A | N/A |
| Cost (In/Out) | $5 / $25 | $2 / $12 |
| Context Window | 1M | 1M |
| Max Output Token | 128k | 128k |
| Input Modality |