Phi-4-multimodal-instruct vs Qwen3.5 397B A17B
Microsoft vs Alibaba — benchmarks, pricing, and capabilities side by side.
- •Qwen3.5 397B A17B has the higher intelligence index (89.3 vs 46.2)
- •Phi-4-multimodal-instruct is cheaper ($0.05 vs $0.39 per 1M input)
- •Qwen3.5 397B A17B is faster
- •Qwen3.5 397B A17B has a larger context window (262K)
| Phi-4-multimodal-instruct | Qwen3.5 397B A17B | |
|---|---|---|
| Intelligence index | 46.2 | 89.3 |
| Developer | Microsoft | Alibaba |
| Type | Multimodal | Multimodal |
| Access | Open weights | Open weights |
| Context window | 128,000 tokens | 262,144 tokens |
| Input price | $0.05 / 1M | $0.39 / 1M |
| Output price | $0.10 / 1M | $2.34 / 1M |
| Speed | 25 tok/s | 53 tok/s |
| Released | February 1, 2025 | February 16, 2026 |
| Parameters | 5600000000 | — |
| Input modalities | — | Text, Image, Video |
| Output modalities | — | Text |
Shared benchmarks
Phi-4-multimodal-instruct
Qwen3.5 397B A17B
GPQA Diamond
31.5
89.3
Humanity’s Last Exam
4.4
27.3