Compare Models
Select models (max 5)
Qwen 3.6 PlusGPT 5.4 Nano
 |  |  |
|---|---|---|
| Legal Research BenchAgentic US legal research | 14.90%±2.48 | 6.25%±1.68 |
| Finance Agent (v2)Core financial analyst tasks | 40.85%±0.13 | 38.22%±1.19 |
| Tax Agent BenchAgentic US corporate tax research | 32.84%±2.75 | 26.58%±2.52 |
| MedCodeMedical billing code support | 36.89%±2.02 | 41.03%±2.26 |
| Code MigrationRewriting programs in new languages | 11.10%±1.31 | 14.47%±4.03 |
| Vibe Code Bench v1.1Building web apps from scratch | 25.57%±3.91 | 26.10%±5.08 |
Benchmarks
Legal Research Bench *
Qwen 3.6 Plus
0.00%± 2.48
(73/73)GPT 5.4 Nano
0.00%± 1.68
(73/73)Finance Agent (v2) *
Qwen 3.6 Plus
0.00%± 0.13
(74/74)GPT 5.4 Nano
0.00%± 1.19
(74/74)Tax Agent Bench *
Qwen 3.6 Plus
0.00%± 2.75
(65/65)GPT 5.4 Nano
0.00%± 2.52
(65/65)MedCode *
Qwen 3.6 Plus
0.00%± 2.02
(104/104)GPT 5.4 Nano
0.00%± 2.26
(104/104)Code Migration *
Qwen 3.6 Plus
0.00%± 1.31
(73/73)GPT 5.4 Nano
0.00%± 4.03
(73/73)Vibe Code Bench v1.1 *
Qwen 3.6 Plus
0.00%± 3.91
(108/108)GPT 5.4 Nano
0.00%± 5.08
(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
33.46%
vs.28.06%
Finance
49.85%
vs.47.21%
Healthcare
56.93%
vs.59.06%
Academic
86.40%
vs.76.09%
Education
44.86%
vs.38.08%
Coding
39.20%
vs.48.59%
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
$0.50
vs.$0.20
Input Cache Write/ 1M Tokens
$0.625
vs.N/A
Input Cache Read/ 1M Tokens
$0.05
vs.$0.02
Output Cost/ 1M Tokens
$3.00
vs.$1.25
Model Metadata
Basic information about each model.
| | |
|---|---|---|
| Model provider | Alibaba | OpenAI |
| Latency | N/A | N/A |
| Cost (In/Out) | $0.5 / $3 | $0.2 / $1.25 |
| Context Window | 984k | 400k |
| Max Output Token | 66k | 128k |
| Input Modality |