← Models
Qwen3.5-397B-A17B
Alibaba Cloud / Qwen Team · open weight · qwen3.5-397b-a17b
Qwen3.5-397B-A17B is Qwen's flagship Mixture-of-Experts model with 397 billion total parameters and 17 billion activated parameters. It delivers state-of-the-art performance across knowledge, reasoning, coding, mathematics, multilingual understanding, instruction following, long context, and agent tasks.
Benchmark scores
| Benchmark | Score |
|---|---|
| MMLU-Redux | 0.95 |
| HMMT 2025 | 0.95 |
| C-Eval | 0.93 |
| HMMT25 | 0.93 |
| IFEval | 0.93 |
| AIME 2026 | 0.91 |
| GPQA | 0.88 |
| MMLU-Pro | 0.88 |
| t2-bench | 0.87 |
| SWE-Bench Verified | 0.76 |
| WideSearch | 0.74 |
| BFCL-V4 | 0.73 |
| Humanity's Last Exam | 0.29 |
Pricing
- DeepInfra$0.45 / $3.00
Input / output per 1M tokens
AA metrics
No Artificial Analysis link yet.
Arena Elo
- No Arena snapshot linked.