Verified · 49/53 providers. Refreshes run at 00:00 and 12:00 UTC. How prices are verified

Model sheet / alibaba

Qwen: Qwen3.5 397B A17B

Languagealibaba/qwen3.5-397b-a17b28 offers · 19 providers

One canonical model, every public offer. Compare input and output separately, check the context attached to each offer, then inspect the raw provider id before integrating.

Raw API ids Public offers onlyPrice method

Provider price range

Complete text-token offers

25 paid complete

Lowest

$1.20

EmpirioLabs AI

Highest

$6.22

LLMTR

Difference

417%

between paid complete offers

Total is input plus output for text tokens. The lowest input or output price can come from a different provider.
Lowest input
$0.17

EmpirioLabs AI

Lowest output
$1.03

EmpirioLabs AI

Lowest total
$1.20

EmpirioLabs AI

Largest reported context
262,144

Lowest input and output prices are independent minima and may come from different providers. “Lowest total” always uses one complete offer.

Benchmarks / source data only

Measured performance

Artificial Analysis and Design Arena scores are published only when OpenRouter reports them for this model. Missing data is never estimated.

Design Arena by category

7 categories · Elo and rank
Show as table
CategoryEloWin %Rank
3D(models)114156.7%#73
Code(models)116852.7%#73
Data viz(models)117153.2%#70
Game dev(models)112950.2%#86
SVG(models)112856.1%#61
UI components(models)115251.4%#84
Website(models)118252.6%#75
License
Not published
Knowledge cutoff
Not published
Collected
Nov 17, 2026

Source: OpenRouter > Artificial Analysis / Design Arena. Scores remain attributed to their originating benchmarks; this page does not average unlike metrics.

Offers / provider sheet

Price by provider

Sort by total, input, output, context or provider. “Total” is input + output for one complete offer; compare the sides separately when your workload is asymmetric.

Browse providers

28 of 28 offers

Specifications

models.dev
Context window
262,144
Max output
0
Input modalities
text, image
Output modalities
text
Reasoning
No
Tool calling
No
Structured output
No
File input
—
Open weights
—
Release date
—
Knowledge cutoff
—
Tokenizer
Qwen3
Family
—
Catalog cost
—

Capabilities

verified specs
Vision