Compare Models
Select models (max 5)
Claude Sonnet 4.6DeepSeek V4 Flash 0731
 |  |  |
|---|---|---|
| Vals IndexGDP-weighted benchmark | N/A | 47.96%±1.24 |
| Legal Research BenchAgentic US legal research | 38.46%±3.38 | 30.29%±3.19 |
| Finance Agent (v2)Core financial analyst tasks | 51.03%±0.33 | 49.52%±0.47 |
| Tax Agent BenchAgentic US corporate tax research | 64.85%±3.11 | 62.56%±3.14 |
| MedCodeMedical billing code support | N/A | 41.41%±2.15 |
| Terminal-Bench ScienceExpert-authored scientific research workflows | N/A | 4.29%±2.44 |
| Code MigrationRewriting programs in new languages | 39.89%±4.14 | 38.63%±4.38 |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | N/A | 18.69%±3.54 |
| Vibe Code Bench v1.1Building web apps from scratch | 51.48%±4.64 | 74.74%±3.44 |
Benchmarks
Vals Index *
Claude Sonnet 4.6
N/ADeepSeek V4 Flash 0731
0.00%± 1.24
(43/43)Legal Research Bench *
Claude Sonnet 4.6
0.00%± 3.38
(73/73)DeepSeek V4 Flash 0731
0.00%± 3.19
(73/73)Finance Agent (v2) *
Claude Sonnet 4.6
0.00%± 0.33
(74/74)DeepSeek V4 Flash 0731
0.00%± 0.47
(74/74)Tax Agent Bench *
Claude Sonnet 4.6
0.00%± 3.11
(65/65)DeepSeek V4 Flash 0731
0.00%± 3.14
(65/65)MedCode *
Claude Sonnet 4.6
N/ADeepSeek V4 Flash 0731
0.00%± 2.15
(104/104)Terminal-Bench Science
Claude Sonnet 4.6
N/ADeepSeek V4 Flash 0731
0.00%± 2.44
(38/38)Code Migration *
Claude Sonnet 4.6
0.00%± 4.14
(73/73)DeepSeek V4 Flash 0731
0.00%± 4.38
(73/73)Terminal-Bench 4.0
Claude Sonnet 4.6
N/ADeepSeek V4 Flash 0731
0.00%± 3.54
(43/43)Vibe Code Bench v1.1 *
Claude Sonnet 4.6
0.00%± 4.64
(108/108)DeepSeek V4 Flash 0731
0.00%± 3.44
(108/108)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, and law
•2026-10-02
Comparison by Industry
Model performance on different sections of the economy.
Legal
41.86%
vs.38.78%
Finance
64.17%
vs.59.94%
Healthcare
N/A
vs.60.89%
Math
N/A
vs.56.00%
Science
N/A
vs.34.37%
Academic
85.51%
vs.88.05%
Education
46.58%
vs.N/A
Coding
50.07%
vs.48.94%
Cyber
N/A
vs.72.08%
Social Mobility
62.45%
vs.N/A
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
N/A
$0.848
Cost / TestVals Index
N/A
vs.$0.848
Input Cost/ 1M Tokens
$3.00
vs.$0.44
Input Cache Write/ 1M Tokens
$3.75
vs.N/A
Input Cache Read/ 1M Tokens
$0.30
vs.$0.014
Output Cost/ 1M Tokens
$15.00
vs.$1.32
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | Anthropic | DeepSeek |
| Latency | N/A | 2585.58s |
| Cost (In/Out) | $3 / $15 | $0.44 / $1.32 |
| Context Window | 1M | 1M |
| Max Output Token | 128k | 384k |
| Input Modality |