family qwen · 21 variants · unknown
| benchmark | category | best score |
|---|---|---|
| QuALITY | knowledge | 100.0 |
| AIME 2025 | math | 92.29 |
| GPQA | reasoning | 86.74 |
| MMLU-Pro | knowledge | 85.85 |
| Terminal-Bench | agentic-coding | 2.0 |
| variant | kind | quant | bits | vRAM | dl | org | year |
|---|---|---|---|---|---|---|---|
| Qwen3.6-35B-A3B-NVFP4 | nvfp4 | nvfp4 | 4.0 | 9,864,413 | nvidia | 2026 | |
| Qwen3.6-35B-A3B-FP8 | fp8 | fp8 | 8.0 | 7,819,997 | Qwen | 2026 | |
| Qwen3.6-35B-A3B | original | bf16 | 16.0 | 6,187,853 | Qwen | 2026 | |
| Qwen3.6-35B-A3B-NVFP4 | nvfp4 | nvfp4 | 4.0 | 870,085 | unsloth | 2026 | |
| Qwen3.6-35B-A3B-GGUF | gguf | gguf | 16.0 | 811,659 | unsloth | 2026 | |
| Qwen3.6-35B-A3B-MTP-GGUF | gguf | gguf | 16.0 | 568,140 | unsloth | 2026 | |
| Qwen3.6-35B-A3B-MLX-4bit | mlx | mlx | 566,422 | lmstudio-community | 2026 | ||
| Qwen3.6-35B-A3B-MLX-8bit | mlx | mlx | 531,128 | lmstudio-community | 2026 | ||
| Qwen3.6-35B-A3B-MLX-6bit | mlx | mlx | 527,346 | lmstudio-community | 2026 | ||
| Qwen3.6-35B-A3B-GGUF | gguf | gguf | 4.0 | 339,981 | lmstudio-community | 2026 | |
| Qwen3.6-35B-A3B-NVFP4-Fast | nvfp4 | nvfp4 | 4.0 | 317,401 | unsloth | 2026 | |
| Qwen_Qwen3.6-35B-A3B-GGUF | gguf | gguf | 1.0 | 97,464 | bartowski | 2026 | |
| Qwen3.6-35B-A3B-GGUF | gguf | gguf | 16.0 | 70,683 | ggml-org | 2026 | |
| Qwen3.6-35B-A3B | original | bf16 | 16.0 | 30,407 | unsloth | 2026 | |
| Qwen3.6-35B-A3B-UD-MLX-4bit | mlx | mlx | 18,560 | unsloth | 2026 | ||
| Qwen3.6-35B-A3B-GGUF | gguf | gguf | 4.0 | 11,547 | mradermacher | 2026 | |
| Qwen3.6-35B-A3B-i1-GGUF | gguf | gguf | 1.0 | 7,336 | mradermacher | 2026 | |
| Qwen3.6-35B-A3B-MLX-8bit | mlx | mlx | 4,392 | unsloth | 2026 | ||
| Qwen3.6-35B-A3B-UD-MLX-3bit | mlx | mlx | 2,589 | unsloth | 2026 | ||
| Holo-3.1-4B-GGUF | gguf | gguf | 16.0 | 1,498 | Mungert | 2026 | |
| Qwen3.6-35B-A3B-MTP-GGUF | gguf | gguf | 16.0 | 1,191 | ggml-org | 2026 |