Qwen3.5 397B A17B vs Qwen3 VL 8B Instruct
Alibaba vs Alibaba — benchmarks, pricing, and capabilities side by side.
- •Qwen3.5 397B A17B has the higher intelligence index (89.3 vs 43)
- •Qwen3 VL 8B Instruct is cheaper ($0.08 vs $0.39 per 1M input)
- •Qwen3 VL 8B Instruct is faster
- •Qwen3.5 397B A17B has a larger context window (262K)
| Qwen3.5 397B A17B | Qwen3 VL 8B Instruct | |
|---|---|---|
| Intelligence index | 89.3 | 43 |
| Developer | Alibaba | Alibaba |
| Type | Multimodal | Multimodal |
| Access | Open weights | Open weights |
| Context window | 262,144 tokens | 256,000 tokens |
| Input price | $0.39 / 1M | $0.08 / 1M |
| Output price | $2.34 / 1M | $0.50 / 1M |
| Speed | 53 tok/s | 145 tok/s |
| Released | February 16, 2026 | October 14, 2025 |
| Parameters | — | — |
| Input modalities | Text, Image, Video | Image, Text |
| Output modalities | Text | Text |
Shared benchmarks
Qwen3.5 397B A17B
Qwen3 VL 8B Instruct
GPQA Diamond
89.3
42.7
Humanity’s Last Exam
27.3
2.9