Qwen3.6 models
16 Qwen3.6 models on Hugging Face, from 386M to 36.0B parameters. At the precision each one is published in, the smallest needs about 0.9 GB of VRAM (Qwen3.6-35B-A3B-DFlash, cheapest live fit: RTX 5060 Ti) and the largest about 80.4 GB (fibo-scene-analyzer, cheapest live fit: RTX PRO 6000). The cheapest way to run Qwen3.6-35B-A3B-DFlash is $0.110/hr.
Qwen3.6 models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| Qwen3.6-35B-A3B-DFlash | 386M | BF16 | 0.9 GB | RTX 5060 Ti | $0.110/hr |
| Bonsai-27B-mlx-1bit | 1.7B | F16 | 3.9 GB | V100 | $0.088/hr |
| Qwen3.8-27B-heretic-dflash | 1.7B | BF16 | 3.9 GB | RTX 5060 Ti | $0.110/hr |
| Qwen3.6-27B-DFlash | 1.7B | BF16 | 3.9 GB | RTX 5060 Ti | $0.110/hr |
| Qwen3.6-27B-NVFP4 | 21.2B | F8_E4M3 | 23.7 GB | RTX 4080 Super | $0.338/hr |
| droplychee-1.0-27b | 27.8B | BF16 | 62.1 GB | A100 | $1.31/hr |
| Qwen3.6-27B | 27.8B | BF16 | 62.1 GB | A100 | $1.31/hr |
| EVE-27b-XENO-HAT-DeepSeek-V4-Flash | 27.8B | BF16 | 62.1 GB | A100 | $1.31/hr |
| Qwen3.6-27B-FP8 | 27.8B | F8_E4M3 | 31.0 GB | RTX 4080 Super | $0.338/hr |
| Qwen3.6-35B-A3B-abliterated-v4 | 34.7B | BF16 | 77.5 GB | A100 | $1.31/hr |
| automotive | 34.7B | BF16 | 77.5 GB | A100 | $1.31/hr |
| Thomson-1.0-Small | 35.1B | BF16 | 78.5 GB | A100 | $1.31/hr |
| fibo-scene-analyzer | 36.0B | BF16 | 80.4 GB | RTX PRO 6000 | $1.38/hr |
| Qwen3.6-35B-A3B | 36.0B | BF16 | 80.4 GB | RTX PRO 6000 | $1.38/hr |
| Qwen3.6-35B-A3B | 36.0B | BF16 | 80.4 GB | RTX PRO 6000 | $1.38/hr |
| Qwen3.6-35B-A3B-FP8 | 36.0B | F8_E4M3 | 40.2 GB | RTX 4090 | $0.441/hr |
VRAM is for the precision the model is published in, with the same overhead and an 8,192-token context assumed on every page; see the methodology. The fit shown is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does.
