LightNav 0 needs about 4.4 GB of memory at Q4_K_M (a 2.7 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 10 quants, from 2.1 GB (Q3_K_S) to 8.9 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.8 GB.
LightNav‑0: Compact embodied navigation model on Qwen3‑VL
LightNav‑0 is a 4.43B‑parameter model that combines vision‑language understanding with navigation. It uses dual‑channel pointing tokens and a residual vector‑quantized action tokenizer to produce 10‑step waypoint trajectories. The model is released under Apache‑2.0.
Good at
- Unified instruction following, object navigation, visual tracking
- Low‑latency inference with llama.cpp
- Compact size for edge deployment
Limits
- No vision support beyond RGB frames
- Requires 8k context for full performance
- No specialized prediction heads
Best for
- Robot navigation in simulated environments
- Real‑robot deployment via WebSocket
- Habitat VLN‑CE and EVT‑Bench evaluation
Who it’s for. Developers building embodied agents that need a single model for navigation, tracking, and instruction following. The model is suitable for research and production use where a lightweight, unified interface is required.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_S | 2.1 GB | 3.8 GB | Fits | Get file → | |
| Q3_K_M | 2.3 GB | 4.0 GB | Fits | Get file → | |
| Q3_K_L | 2.4 GB | 4.1 GB | Fits | Get file → | |
| Q4_0 | 2.6 GB | 4.3 GB | Fits | Get file → | |
| Q4_K_S | 2.6 GB | 4.3 GB | Fits | Get file → | |
| Q4_K_M | 2.7 GB | 4.4 GB | Fits | Get file → | |
| Q5_0 | 3.1 GB | 4.8 GB | Fits | Get file → | |
| Q5_K_S | 3.1 GB | 4.8 GB | Fits | Get file → | |
| Q5_K_M | 3.2 GB | 4.9 GB | Fits | Get file → | |
| BF16 | 8.9 GB | 10.6 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (2.73 GB file, ~4.4 GB memory). For 12 GB VRAM, Q5_K_M (3.17 GB file, ~4.9 GB memory) offers higher quality. The BF16 variant (8.87 GB file, ~10.6 GB memory) provides full precision if memory permits.
What GPU do you need for LightNav 0?
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 Q3_K_M (2.3 GB file, about 4.0 GB in use).
- 6 GB VRAM: the best quant that fits is Q5_K_M (3.2 GB file, about 4.9 GB in use).
- 8 GB VRAM: the best quant that fits is Q5_K_M (3.2 GB file, about 4.9 GB in use).
- 12 GB VRAM: the best quant that fits is BF16 (8.9 GB file, about 10.6 GB in use).
Which LightNav 0 quant should you choose?
Q4_K_M is the usual starting point for LightNav 0: it balances file size and quality well. Here it is a 2.7 GB file.
If memory is tight, the smallest file is Q3_K_S at 2.1 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about LightNav 0
How much VRAM does LightNav 0 need?
About 4.4 GB at Q4_K_M with an 8K context window. That is the 2.7 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 LightNav 0 on an 8 GB GPU?
Yes, with Q5_K_M: the 3.2 GB file needs about 4.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of LightNav 0 GGUF?
It depends on the quant: Q3_K_S is 2.1 GB, Q3_K_M is 2.3 GB, Q3_K_L is 2.4 GB, Q4_0 is 2.6 GB, Q4_K_S is 2.6 GB, Q4_K_M is 2.7 GB and 4 more.
What context length does LightNav 0 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.2 GB at 8K and 4.8 GB at 32K.
Where can I download LightNav 0 GGUF?
The files are on Hugging Face in the prithivMLmods/LightNav-0-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 LightNav 0 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 LightNav‑0 on 8 GB VRAM?
Yes, the Q4_K_M quantized model requires about 4.4 GB of VRAM at 8k context, which fits within an 8 GB GPU.
Does LightNav‑0 support vision input?
LightNav‑0 accepts a SlowFast‑compressed history of first‑person RGB frames (256×448 at 4 fps) along with a natural‑language instruction.
What license applies to LightNav‑0?
LightNav‑0 is released under the Apache License 2.0.