ModelWatch

Compare models

Up to 4 models side by side.

Provider
Alibaba (Qwen)
Listed
Date added to OpenRouter
Feb 16, 2026
Input price
USD per 1M tokens
$0.45
Output price
USD per 1M tokens
$3
Workload cost
Chat reply (2K in, 500 out)
$0.0024
Context window
262K
Max output
82K
Output speed
tokens per second
53
Time to first token
seconds
2.47s
Intelligence Index
Artificial Analysis Intelligence Index: a composite of reasoning, knowledge, math and coding evaluations.
n/a
Engineering Index
Artificial Analysis Engineering Index: engineering and technical problem solving.
n/a
Inputs
text, image, video
Open weights
Yes
Good at
visionopen-weights
Notes
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...

Best value in each row is highlighted. Share this page's URL to share the comparison.