family qwen · 19 variants · unknown
| benchmark | category | best score |
|---|---|---|
| MMLU | knowledge | 89.5 |
| AIME 2024 | math | 80.4 |
| GPQA | reasoning | 65.8 |
| variant | kind | quant | bits | vRAM | dl | org | year |
|---|---|---|---|---|---|---|---|
| Qwen3-30B-A3B-abliterated | abliterated | abliterated | 454,663 | mlabonne | 2025 | ||
| Qwen3-30B-A3B-FP8 | fp8 | fp8 | 8.0 | 175,061 | Qwen | 2025 | |
| Qwen3-30B-A3B-GPTQ-Int4 | gptq | gptq | 147,555 | Qwen | 2025 | ||
| Qwen3-30B-A3B-GGUF | gguf | gguf | 4.0 | 95,301 | Qwen | 2025 | |
| Qwen3-30B-A3B-NVFP4 | nvfp4 | nvfp4 | 4.0 | 62,746 | nvidia | 2025 | |
| Qwen3-30B-A3B-GGUF | gguf | gguf | 16.0 | 56,540 | unsloth | 2025 | |
| Qwen_Qwen3-30B-A3B-GGUF | gguf | gguf | 2.0 | 11,204 | bartowski | 2025 | |
| Qwen3-30B-A3B-MLX-4bit | mlx | mlx | 10,375 | lmstudio-community | 2025 | ||
| Qwen3-30B-A3B-MLX-8bit | mlx | mlx | 10,014 | lmstudio-community | 2025 | ||
| Qwen3-30B-A3B-GGUF | gguf | gguf | 3.0 | 8,218 | lmstudio-community | 2025 | |
| Qwen3-30B-A3B-128K-GGUF | gguf | gguf | 16.0 | 2,747 | unsloth | 2025 | |
| Qwen3-30B-A3B | original | bf16 | 16.0 | 2,454 | unsloth | 2025 | |
| Qwen3-30B-A3B-bnb-4bit | bnb-4bit | 4bit | 864 | unsloth | 2025 | ||
| Qwen3-30B-A3B-MLX-4bit | mlx | mlx | 550 | Qwen | 2025 | ||
| Qwen3-30B-A3B-GGUF | gguf | gguf | 4.0 | 279 | ggml-org | 2025 | |
| Qwen3-30B-A3B-MLX-8bit | mlx | mlx | 139 | Qwen | 2025 | ||
| Qwen3-30B-A3B-MLX-6bit | mlx | mlx | 126 | Qwen | 2025 | ||
| Qwen3-30B-A3B-MLX-bf16 | mlx | mlx | 120 | Qwen | 2025 | ||
| Qwen3-30B-A3B-ArliAI-RpR-v4-Fast | original | bf16 | 16.0 | 78 | ArliAI | 2025 |