Granite 3.3 8B Instruct vs Qwen3.5 397B A17B
IBM vs Alibaba — benchmarks, pricing, and capabilities side by side.
- •Qwen3.5 397B A17B has the higher intelligence index (89.3 vs 68.5)
| Granite 3.3 8B Instruct | Qwen3.5 397B A17B | |
|---|---|---|
| Intelligence index | 68.5 | 89.3 |
| Developer | IBM | Alibaba |
| Type | Multimodal | Multimodal |
| Access | Open weights | Open weights |
| Context window | — | 262,144 tokens |
| Input price | — | $0.39 / 1M |
| Output price | — | $2.34 / 1M |
| Speed | — | 53 tok/s |
| Released | April 16, 2025 | February 16, 2026 |
| Parameters | 8000000000 | — |
| Input modalities | — | Text, Image, Video |
| Output modalities | — | Text |