About
The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. Compared to the...
Related Models
Pricing
Input
$0.07
per 1M tokens
Output
$0.26
per 1M tokens
Blended
$0.11
per 1M tokens
Cheaper than 82% of models. Median price is $0.67/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.11
Monthly
$3.41
vs. Similar Models
Qwen: Qwen3.5-9B
$0.11-1%
Meta: Llama 3.2 3B Instruct
$0.12+5%
Qwen: Qwen3 Coder 30B A3B Instruct
$0.12+5%
Granite 4.0 H Small
$0.11-6%
Performance
Context Window
1.0M
tokens
Larger than 80% of models
Max Output
66K
tokens
7% of context
Context Window Comparison
Anthropic: Claude Fable 5
1.0MSame
Anthropic: Claude Opus 4.8
1.0MSame
Anthropic: Claude Opus 4.7
1.0MSame