Internvl3-38B needs about 22.5 GB of memory at Q4_K_M (a 19.8 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 24 quants, from 7.6 GB (IQ1_S) to 65.5 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.
InternVL3‑38B: 32.8B multimodal LLM with vision, coding, and tool use
InternVL3‑38B is a 32.8B‑parameter multimodal large language model. It supports text, vision, coding, and tool usage. The model is trained on 32768‑token context windows and released under the Apache‑2.0 license.
Good at
- Vision‑guided reasoning
- Tool‑enabled coding
- Multimodal dialogue
Limits
- Requires large VRAM (minimum 10.3 GB for IQ1_S)
- No native vision support in llama.cpp without custom adapters
- No explicit language list provided
Best for
- Image‑captioning with code snippets
- GUI automation with tool calls
- Complex multimodal Q&A
Who it’s for. Developers and researchers needing a versatile multimodal model that can process images, write code, and invoke external tools. The model is suitable for building advanced chatbots, visual assistants, and automated analysis pipelines.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 7.6 GB | 10.3 GB | Too big | Get file → | |
| IQ1_M | 8.2 GB | 10.9 GB | Too big | Get file → | |
| IQ2_XXS | 9.3 GB | 11.9 GB | Too big | Get file → | |
| IQ2_M | 11.4 GB | 14.1 GB | Too big | Get file → | |
| Q2_K | 12.3 GB | 15.0 GB | Too big | Get file → | |
| Q2_K_L | 12.5 GB | 15.1 GB | Too big | Get file → | |
| Q2_K_XL | 12.7 GB | 15.4 GB | Too big | Get file → | |
| IQ3_XXS | 13.0 GB | 15.6 GB | Too big | Get file → | |
| Q3_K_S | 14.4 GB | 17.0 GB | Too big | Get file → | |
| Q3_K_M | 15.9 GB | 18.6 GB | Too big | Get file → | |
| Q3_K_XL | 16.3 GB | 19.0 GB | Too big | Get file → | |
| IQ4_XS | 17.7 GB | 20.4 GB | Too big | Get file → | |
| IQ4_NL | 18.7 GB | 21.3 GB | Too big | Get file → | |
| Q4_K_S | 18.8 GB | 21.4 GB | Too big | Get file → | |
| Q4_K_M | 19.8 GB | 22.5 GB | Too big | Get file → | |
| Q4_K_XL | 20.1 GB | 22.7 GB | Too big | Get file → | |
| Q4_1 | 20.6 GB | 23.3 GB | Too big | Get file → | |
| Q5_K_S | 22.6 GB | 25.3 GB | Too big | Get file → | |
| Q5_K_M | 23.3 GB | 25.9 GB | Too big | Get file → | |
| Q5_K_XL | 23.3 GB | 25.9 GB | Too big | Get file → | |
| Q6_K | 26.9 GB | 29.5 GB | Too big | Get file → | |
| Q6_K_XL | 28.7 GB | 31.3 GB | Too big | Get file → | |
| Q8_K_XL | 39.2 GB | 41.8 GB | Too big | Get file → | |
| BF16 | 65.5 GB (2 parts) | 68.2 GB | Too big | Get file → |
For 8 GB VRAM, the smallest usable quant is IQ1_S (file size 7.63 GB, estimated 10.3 GB memory). For 12 GB VRAM, IQ1_M (8.24 GB, 10.9 GB memory) is a good balance. For 24 GB VRAM, Q4_K_M (19.85 GB, 22.5 GB memory) offers higher quality. Estimates are approximate.
What GPU do you need for Internvl3-38B?
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: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: the best quant that fits is IQ2_XXS (9.3 GB file, about 11.9 GB in use).
- 16 GB VRAM: the best quant that fits is IQ3_XXS (13.0 GB file, about 15.6 GB in use).
- 24 GB VRAM: the best quant that fits is Q4_1 (20.6 GB file, about 23.3 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K_XL (28.7 GB file, about 31.3 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_K_XL (39.2 GB file, about 41.8 GB in use).
Which Internvl3-38B quant should you choose?
Q4_K_M is the usual starting point for Internvl3-38B: it balances file size and quality well. Here it is a 19.8 GB file.
If memory is tight, the smallest file is IQ1_S at 7.6 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Internvl3-38B
How much VRAM does Internvl3-38B need?
About 22.5 GB at Q4_K_M with an 8K context window. That is the 19.8 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 Internvl3-38B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 10.3 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Internvl3-38B GGUF?
It depends on the quant: IQ1_S is 7.6 GB, IQ1_M is 8.2 GB, IQ2_XXS is 9.3 GB, IQ2_M is 11.4 GB, Q2_K is 12.3 GB, Q2_K_L is 12.5 GB and 18 more.
What context length does Internvl3-38B support?
The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.
Where can I download Internvl3-38B GGUF?
The files are on Hugging Face in the unsloth/InternVL3-38B-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 Internvl3-38B 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 InternVL3‑38B on 8 GB VRAM?
No. The smallest quant requires about 10.3 GB of VRAM at 8k context.
Does InternVL3‑38B support tool usage?
Yes. The model detects tool‑usage ability and can invoke external tools.
Is InternVL3‑38B available for commercial use?
It is released under the Apache‑2.0 license, which permits commercial use.