unsloth
Qwen3.5 2B
Qwen3.5 2B is a 1.88B dense qwen35 model from unsloth. Configured context length is 262,144 tokens. License: apache-2.0. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters1.9B
Context256K
Attentionhybrid
Licenseapache-2.0
24 layers · 2 KV heads · head dim 256 · 18 linear-attention layers · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 3.52 GB | Hugging Face file bytes |
| UD-Q8_K_XL | 2.64 GB | Hugging Face file bytes |
| Q8_0 | 1.87 GB | Hugging Face file bytes |
| UD-Q6_K_XL | 1.74 GB | Hugging Face file bytes |
| Q6_K | 1.47 GB | Hugging Face file bytes |
| UD-Q5_K_XL | 1.37 GB | Hugging Face file bytes |
| Q5_K_M | 1.34 GB | Hugging Face file bytes |
| Q5_K_S | 1.29 GB | Hugging Face file bytes |
| UD-Q4_K_XL | 1.25 GB | Hugging Face file bytes |
| Q4_K_M | 1.19 GB | Hugging Face file bytes |
| Q4_K_S | 1.13 GB | Hugging Face file bytes |
| IQ4_NL | 1.13 GB | Hugging Face file bytes |
| IQ4_XS | 1.09 GB | Hugging Face file bytes |
| Q4_1 | 1.20 GB | Hugging Face file bytes |
| Q4_0 | 1.13 GB | Hugging Face file bytes |
| UD-Q3_K_XL | 1.08 GB | Hugging Face file bytes |
| Q3_K_M | 1.03 GB | Hugging Face file bytes |
| Q3_K_S | 0.96 GB | Hugging Face file bytes |
| UD-IQ3_XXS | 0.87 GB | Hugging Face file bytes |
| UD-Q2_K_XL | 0.90 GB | Hugging Face file bytes |
| UD-IQ2_XXS | 0.72 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Runs great | BF16 | 41–56 |
| GeForce RTX 3060 12GB | Runs great | BF16 | 51–70 |
| GeForce RTX 4060 Ti 16GB | Runs great | BF16 | 41–56 |
| GeForce RTX 4090 | Runs great | BF16 | 143–195 |
| GeForce RTX 5090 | Runs great | BF16 | 254–347 |
| Apple M4 Pro | Runs great | BF16 | 39–53 |
| Apple M5 | Runs great | BF16 | 22–30 |
| CPU only · DDR5-5600 | Runs on CPU | BF16 | 13–17 |
Smallest devices that fit it fully
- GeForce RTX 3050 6GB — Runs, BF16, 24–33 tok/s
- GeForce RTX 2060 6GB — Runs, BF16, 48–65 tok/s
- GeForce GTX 1660 Super — Runs, BF16, 48–65 tok/s
- GeForce GTX 1660 Ti — Runs, BF16, 41–56 tok/s
- Intel Arc A380 — Runs, BF16, 26–36 tok/s
- GeForce RTX 5060 Ti 8GB — Runs great, BF16, 64–87 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (BF16).
llama.cpp
llama-server -hf unsloth/Qwen3.5-2B-GGUF:BF16 -c 8192 -ngl 24
Ollama
ollama run qwen3.5
LM Studio
lms get unsloth/Qwen3.5-2B-GGUF # Then load quant BF16