LocateAnything 3B — Hardware Requirements & GPU Compatibility
Vision
LocateAnything-3B is NVIDIA's 3.8-billion-parameter vision-language model for visual grounding rather than open-ended chat: referring-expression grounding, dense multi-object detection, GUI element grounding, point-based localization, and document or OCR layout grounding. Its core contribution is Parallel Box Decoding, which predicts a complete bounding box in one parallel step instead of token-by-token autoregressive decoding, giving up to 2.5x higher throughput while preserving geometric consistency. It combines a Qwen2.5-3B-Instruct language backbone with a MoonViT-SO-400M vision encoder and was trained on 12 million images with over 138 million grounding queries; NVIDIA has folded it into the Nemotron 3 Nano Omni model for agentic and computer-use grounding. At under 4 billion parameters, it runs on a single consumer GPU. Context length is 32,768 tokens. It is released under NVIDIA's non-commercial research license, permitting academic and non-profit use only; commercial use requires a separate license from NVIDIA. It was published in May 2026.
87.4K downloads
3.0K likesSpecifications
- Publisher
- NVIDIA
- Parameters
- 3.8B
- Architecture
- LocateAnythingForConditionalGeneration
- Context Length
- 32,768 tokens
- Vocabulary Size
- 152,681
- Release Date
- 2026-03-02
- License
- Other
Get Started
HuggingFace
How Much VRAM Does LocateAnything 3B Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 2 GB | 3.1 GB | 1.63 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 2.0 GB | 3.2 GB | 1.68 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 2.2 GB | 3.4 GB | 1.87 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 2.3 GB | 3.4 GB | 1.92 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 2.7 GB | 3.8 GB | 2.30 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 3.1 GB | 4.2 GB | 2.73 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 3.5 GB | 4.7 GB | 3.16 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 4.2 GB | 5.3 GB | 3.83 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run LocateAnything 3B?
Q4_K_M · 2.7 GBLocateAnything 3B (Q4_K_M) requires 2.7 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 4+ GB is recommended. Using the full 33K context window can add up to 1.1 GB, bringing total usage to 3.8 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run LocateAnything 3B?
Q4_K_M · 2.7 GB59 devices with unified memory can run LocateAnything 3B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomWhere to Download LocateAnything 3B
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does LocateAnything 3B need?
LocateAnything 3B requires 2.7 GB of VRAM at Q4_K_M, or 8.0 GB at BF16. Full 33K context adds up to 1.1 GB (3.8 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 3.8B × 4.8 bits ÷ 8 = 2.3 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 1.5 GB (at full 33K context)
VRAM usage by quantization
Q4_K_M2.7 GBQ4_K_M + full context3.8 GB- What's the best quantization for LocateAnything 3B?
For LocateAnything 3B, Q4_K_M (2.7 GB) offers the best balance of quality and VRAM usage. Q5_K_S (3.0 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 2 GB.
VRAM requirement by quantization
Q2_K2.0 GBQ4_02.3 GBQ4_K_S2.5 GBQ4_K_M ★2.7 GBQ5_K_M3.1 GBBF168.0 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run LocateAnything 3B on a Mac?
LocateAnything 3B requires at least 2 GB at Q2_K, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run LocateAnything 3B locally?
Yes — LocateAnything 3B can run locally on consumer hardware. At Q4_K_M quantization it needs 2.7 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is LocateAnything 3B?
At Q4_K_M, LocateAnything 3B can reach ~1798 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~245 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 2.7 × 0.65 = ~1948 tok/s
Estimated speed at Q4_K_M (2.7 GB)
~1948 tok/s~245 tok/s~1948 tok/s~1798 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of LocateAnything 3B?
At Q4_K_M, the download is about 2.30 GB. The full-precision BF16 version is 7.66 GB. The smallest option (Q2_K) is 1.63 GB.
- Which GPUs can run LocateAnything 3B?
52 consumer GPUs can run LocateAnything 3B at Q4_K_M (2.7 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
- Which devices can run LocateAnything 3B?
59 devices with unified memory can run LocateAnything 3B at Q4_K_M (2.7 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Apple iPhone 17 Pro, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.