qwen/qwen3-30b-a3b Details | |
|---|---|
| Overview | |
| Description | Alibaba's efficient Mixture-of-Experts model with 30B total parameters and only 3B activated per token, optimized for fast inference. |
| Author | Qwen |
| Category | Text Generation |
| Context length | 131K |
| Providers | 3 |
| Pricing (per 1M tokens) | |
| Input | $0 |
| Output | $0 |
| Capabilities | |
| Reasoning | |
| Tool calling | |
| Vision | – |
| Streaming | |
| Token activity (30 days) | |
| Usage | |
| Artificial Analysis | |
| Intelligence | 12.5 |
| Coding | 13.3 |
| Benchmarks | |
| GPQA Diamond | 0.5 |
| HLE | 0.0 |
| MMLU-Pro | 0.7 |
| IFEval | 0.3 |
| LiveCodeBench | 0.3 |
| AIME | 0.3 |
| MATH-500 | 0.9 |