CapRL Qwen3VL 2B needs about 2.7 GB of memory at Q4_K_M (a 1.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 37 quants, from 0.0 GB to 4.1 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 3.8 GB.
CapRL Qwen3VL 2B – 2.03B parameter vision‑language model
CapRL Qwen3VL 2B is a 2.03B parameter model fine‑tuned from Qwen3‑VL‑2B. It supports vision and tools and is optimized for image captioning. The model is released under the apache‑2.0 license.
Good at
- High‑quality image captioning
- Vision‑language reasoning
- Tool‑enabled interactions
Limits
- Requires 8‑k context for memory estimates
- No explicit vision‑only mode beyond image captioning
- No mention of audio or video support
Best for
- Captioning charts and infographics
- Generating concise image descriptions
- Integrating with vLLM APIs for lightweight deployment
Who it’s for. Developers and researchers needing a compact vision‑language model for captioning tasks. The model is suitable for environments with limited GPU memory while still providing strong visual understanding.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 1.4 GB | Fits | Get file → | |
| IQ1_S | 0.6 GB | 2.1 GB | Fits | Get file → | |
| IQ1_M | 0.6 GB | 2.1 GB | Fits | Get file → | |
| IQ2_XXS | 0.7 GB | 2.1 GB | Fits | Get file → | |
| IQ2_XS | 0.7 GB | 2.2 GB | Fits | Get file → | |
| IQ2_S | 0.8 GB | 2.2 GB | Fits | Get file → | |
| IQ2_M | 0.8 GB | 2.3 GB | Fits | Get file → | |
| Q2_K_S | 0.8 GB | 2.3 GB | Fits | Get file → | |
| Q2_K | 0.9 GB | 2.3 GB | Fits | Get file → | |
| Q2_K | 0.9 GB | 2.3 GB | Fits | Get file → | |
| IQ3_XXS | 0.9 GB | 2.3 GB | Fits | Get file → | |
| IQ3_XS | 1.0 GB | 2.4 GB | Fits | Get file → | |
| Q3_K_S | 1.0 GB | 2.4 GB | Fits | Get file → | |
| IQ3_S | 1.0 GB | 2.4 GB | Fits | Get file → | |
| Q3_K_S | 1.0 GB | 2.4 GB | Fits | Get file → | |
| IQ3_M | 1.0 GB | 2.5 GB | Fits | Get file → | |
| Q3_K_M | 1.1 GB | 2.5 GB | Fits | Get file → | |
| Q3_K_M | 1.1 GB | 2.5 GB | Fits | Get file → | |
| Q3_K_L | 1.1 GB | 2.6 GB | Fits | Get file → | |
| Q3_K_L | 1.1 GB | 2.6 GB | Fits | Get file → | |
| IQ4_XS | 1.2 GB | 2.6 GB | Fits | Get file → | |
| IQ4_XS | 1.2 GB | 2.6 GB | Fits | Get file → | |
| IQ4_NL | 1.2 GB | 2.7 GB | Fits | Get file → | |
| Q4_0 | 1.2 GB | 2.7 GB | Fits | Get file → | |
| Q4_K_S | 1.2 GB | 2.7 GB | Fits | Get file → | |
| Q4_K_S | 1.2 GB | 2.7 GB | Fits | Get file → | |
| Q4_K_M | 1.3 GB | 2.7 GB | Fits | Get file → | |
| Q4_K_M | 1.3 GB | 2.7 GB | Fits | Get file → | |
| Q4_1 | 1.3 GB | 2.8 GB | Fits | Get file → | |
| Q5_K_S | 1.4 GB | 2.9 GB | Fits | Get file → | |
| Q5_K_S | 1.4 GB | 2.9 GB | Fits | Get file → | |
| Q5_K_M | 1.5 GB | 2.9 GB | Fits | Get file → | |
| Q5_K_M | 1.5 GB | 2.9 GB | Fits | Get file → | |
| Q6_K | 1.7 GB | 3.1 GB | Fits | Get file → | |
| Q6_K | 1.7 GB | 3.1 GB | Fits | Get file → | |
| Q8_0 | 2.2 GB | 3.6 GB | Fits | Get file → | |
| F16 | 4.1 GB | 5.5 GB | Fits | Get file → |
For 8‑GB VRAM, use IQ1_S (0.62 GB, 2.1 GB memory) or Q2_K_S (0.84 GB, 2.3 GB memory). For 16‑GB VRAM, Q3_K_M (1.07 GB, 2.5 GB memory) or Q4_K_M (1.28 GB, 2.7 GB memory) offer higher fidelity. Larger GPUs can handle Q5_K_M (1.47 GB, 2.9 GB memory) or Q6_K (1.67 GB, 3.1 GB memory).
What GPU do you need for CapRL Qwen3VL 2B?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is Q8_0 (2.2 GB file, about 3.6 GB in use).
- 6 GB VRAM: the best quant that fits is F16 (4.1 GB file, about 5.5 GB in use).
Which CapRL Qwen3VL 2B quant should you choose?
Q4_K_M is the usual starting point for CapRL Qwen3VL 2B: it balances file size and quality well. Here it is a 1.3 GB file.
The Q8_0 file is 2.2 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about CapRL Qwen3VL 2B
How much VRAM does CapRL Qwen3VL 2B need?
About 2.7 GB at Q4_K_M with an 8K context window. That is the 1.3 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run CapRL Qwen3VL 2B on an 8 GB GPU?
Yes, with F16: the 4.1 GB file needs about 5.5 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of CapRL Qwen3VL 2B GGUF?
It depends on the quant: file is 0.0 GB, IQ1_S is 0.6 GB, IQ1_M is 0.6 GB, IQ2_XXS is 0.7 GB, IQ2_XS is 0.7 GB, IQ2_S is 0.8 GB and 31 more.
What context length does CapRL Qwen3VL 2B support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.9 GB at 8K and 3.8 GB at 32K.
Where can I download CapRL Qwen3VL 2B GGUF?
The files are on Hugging Face in the prithivMLmods/CapRL-Qwen3VL-2B-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does CapRL Qwen3VL 2B use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run CapRL Qwen3VL 2B on 8 GB VRAM?
Yes, using IQ1_S or Q2_K_S quantizations keeps memory usage below 8 GB, with approximate 2.1–2.3 GB VRAM usage at 8‑k context.
Does the model support image inputs in base64 format?
Yes, the model accepts base64‑encoded images for captioning and visual reasoning tasks.
Is there a license restriction for commercial use?
The model is licensed under apache‑2.0, which permits commercial use with attribution.