Compare Models
Select models (max 7)
DeepSeek V4 Pro 0813Claude Fable 5.2
 |  |  |
|---|---|---|
| Vals IndexGDP-weighted benchmark | 52.97%±1.04 | 71.31%±0.96 |
| Legal Research BenchAgentic US legal research | 40.87%±3.42 | 46.15% ±3.46 |
| Finance Agent v2.1Core financial analyst tasks | 57.12%±2.08 | 69.47%±1.98 |
| Tax Agent BenchAgentic US corporate tax research | 26.76%±3.09 | 32.09%±3.37 |
| MedCodeMedical billing code support | 42.47%±2.16 | N/A |
| Terminal-Bench-ScienceExpert-authored scientific research workflows | 4.29%±2.44 | 55.71%±5.98 |
| Code Migration v1.1Rewriting programs in new languages | 51.99%±2.23 | 83.27%±3.12 |
| Terminal-Bench 4.0Frontier-difficulty terminal tasks | 14.14%±2.02 | 65.15%±2.31 |
| Vibe Code Bench v1.1Building web apps from scratch | 82.30%±3.15 | 93.77%±1.15 |
Benchmarks
Vals Index *
DeepSeek V4 Pro 0813
0.00%± 1.04
(36/36)Claude Fable 5.2
0.00%± 0.96
(36/36)Legal Research Bench *
DeepSeek V4 Pro 0813
0.00%± 3.42
(93/93)Claude Fable 5.2
0.00%± 3.46
(93/93)Finance Agent v2.1 *
DeepSeek V4 Pro 0813
0.00%± 2.08
(39/39)Claude Fable 5.2
0.00%± 1.98
(39/39)Tax Agent Bench *
DeepSeek V4 Pro 0813
0.00%± 3.09
(84/84)Claude Fable 5.2
0.00%± 3.37
(84/84)MedCode *
DeepSeek V4 Pro 0813
0.00%± 2.16
(107/107)Claude Fable 5.2
N/ATerminal-Bench-Science
DeepSeek V4 Pro 0813
0.00%± 2.44
(45/45)Claude Fable 5.2
0.00%± 5.98
(45/45)Code Migration v1.1 *
DeepSeek V4 Pro 0813
0.00%± 2.23
(30/30)Claude Fable 5.2
0.00%± 3.12
(30/30)Terminal-Bench 4.0
DeepSeek V4 Pro 0813
0.00%± 2.02
(63/63)Claude Fable 5.2
0.00%± 2.31
(63/63)Vibe Code Bench v1.1 *
DeepSeek V4 Pro 0813
0.00%± 3.15
(127/127)Claude Fable 5.2
0.00%± 1.15
(127/127)Overall performance
Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, legal, and tax
•2026-11-09
Comparison by Industry
Model performance on different sections of the economy.
Legal
48.93%
vs.41.71%
Finance
55.51%
vs.59.65%
Healthcare
61.32%
vs.N/A
Math
50.00%
vs.100.00%
Science
4.29%
vs.63.74%
Academic
89.70%
vs.N/A
Coding
50.59%
vs.64.70%
Cyber
37.38%
vs.N/A
Cost Analysis
Per task and per token model pricing.
Vals Index · USD
$2.50
$17.94
Cost / TestVals Index
$2.50
vs.$17.94
Input Cost/ 1M Tokens
$1.32
vs.$8.00
Input Cache Write/ 1M Tokens
N/A
vs.$10.00
Input Cache Read/ 1M Tokens
$0.044
vs.$0.20
Output Cost/ 1M Tokens
$3.96
vs.$40.00
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | DeepSeek | Anthropic |
| Latency | 3482.91s | 2621.82s |
| Cost (In/Out) | $1.32 / $3.96 | $8 / $40 |
| Context Window | 1M | 1M |
| Max Output Token | 384k | 128k |
| Input Modality |