Loading...
Loading...
The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall performance is comparable to that of the Qwen3.5-27B.
Quality Index
37.1
60th of 442
Top 14%
Coding Index
30.3
78th of 352
Top 22%
Price/1M
$0.69
418th cheapest
122% above median
Top 62%
Speed
124 tok/s
Top 22%
TTFT
1.05s
Context Window
262K
61st largest
Top 25%
Input
$0.25
per 1M tokens
Output
$2.00
per 1M tokens
Blended
$0.69
per 1M tokens
Cheaper than 38% of models. Median price is $0.31/1M tokens.
Daily
$0.69
Monthly
$20.64
124
tokens/sec
Faster than 78% of models
1.05
seconds
Faster than 27% of models
17.22
seconds
Faster than 15% of models
Market Median
46 tok/s
171% faster
Median TTFT
0.42s
150% slower
Throughput/Dollar
180
tok/s per $/1M
Speed Comparison
Context Window
262K
tokens
Larger than 75% of models
Max Output
66K
tokens
25% of context
2.4M
1.2K
48-80 GB
A100 80GB