DeepSeek OCR — Hardware Requirements & GPU Compatibility
Vision
DeepSeek OCR is a 3.3-billion-parameter vision-language model from DeepSeek, purpose-built for optical character recognition and document parsing rather than general chat. It pairs a vision encoder with a Mixture-of-Experts decoder that activates roughly 1.1 billion parameters per token, keeping decoding fast while all expert weights still need to fit in memory. Its core idea is compressing a page of text into a much smaller set of image tokens before decoding, and it is small enough to run on a single consumer GPU once quantized. Context length is limited to 8,192 tokens, reflecting its page-oriented use case. It is released under the MIT license, a highly permissive option for commercial use, and was published in October 2025, introducing "optical context compression" to shrink the token count needed for OCR.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 8K)Best on 24 GB (Q4_K_M · ~519 tok/s · 8K)Macs from 8 GB (Q4_K_M · ~72 tok/s)
MIT
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run DeepSeek OCR
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/sabafallah/DeepSeek-OCR-GGUF:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 21.2 GB free · 8K context
Download: deepseek-ocr-Q4_K_M.gguf1.95 GBsabafallah
Ways to runOther ways
each opens step-by-stepRun in cloud
Your own private copy · Q4_K_M · 24 GB GPU
- The easiest way to try this model
- Running in a few minutes, easy setup
- Chat right here or connect any app
$5+ bonus for new users
Runs on RTX 3060 12GB (9 GB headroom) · Q4_K_M
- Generation speed
- ~237 tok/s
- generation speed
- Cost per 1M output tokens
- $0.072
- per 1M output tokens
How Much VRAM Does DeepSeek OCR Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 1.8 GB | — | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 8K context · ~617 tok/s | |
| 2.0 GB | — | 3-bit medium — 3-bit medium quantization | Fits: 8K context · ~578 tok/s | |
| 2.1 GB | — | 4-bit · legacy — 4-bit legacy quantization | Fits: 8K context · ~571 tok/s | |
| 2.4 GB | — | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 8K context · ~519 tok/s | |
| 2.5 GB | — | 5-bit · legacy — 5-bit legacy quantization | Fits: 8K context · ~508 tok/s | |
| 2.8 GB | — | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 8K context · ~471 tok/s | |
| 3.2 GB | — | 6-bit, very good — 6-bit quantization, very good quality | Fits: 8K context · ~431 tok/s | |
| 3.8 GB | — | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 8K context · ~381 tok/s | |
| 7.1 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 8K context · ~229 tok/s |
Selected Q4_K_M
4-bit medium quantization2.8GBat 8K context
- Weights
- 2.0 GB
- KV cache
- 0.5 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 21.2 GB free · ~519 tok/s
≥50 tok/s needs ≥66 GB/s · RTX 5090 ~728, RTX 3090 Ti ~519
Which Hardware Can Run DeepSeek OCR?
Q4_K_M · 2.4 GBDeepSeek OCR (Q4_K_M) requires 2.4 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 8K context window can add up to 0.4 GB, bringing total usage to 2.8 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run DeepSeek OCR, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 5.2 GB free · up to 8K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~427Simulate output on RTX 3080RTX 2080 Ti~366Simulate output on RTX 2080 TiRTX 3070 Ti~362Simulate output on RTX 3070 TiM3*8 GB~72Simulate output on iMac 24" M3 (8 GB)Orin Nano~76Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~51Simulate output on iPhone 1720 GB
fits · 17.2 GB free · up to 8K ctx
1 GPU
RX 7900 XT~242Simulate output on RX 7900 XT32–40 GB
fits · 29.2 GB free · up to 8K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~728Simulate output on RTX 5090M5 Max36 GB~201Simulate output on Mac Studio M5 Max (36 GB)M532 GB~99Simulate output on MacBook Air 13" M5 (32 GB)Orin~145Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~94Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to DeepSeek OCR
DeepSeek family · 5 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Unlimited OCR 2Baidu | — | 3.3 GB | ~517 | same size |
| D1 3BLiquid AI | — | 2.6 GB | ~285 | same size |
| Jina Ocr V1jinaai | — | 3.3 GB | ~509 | same size |
| Unlimited OCRBaidu | — | 3.3 GB | ~519 | same size |
| Granite 4.0 MicroIBM | — | 3.7 GB | ~261 | same size |
| Llama 3.2 3B InstructMeta | — | 4.1 GB | ~309 | same size |
Specifications
- Publisher
- DeepSeek
- Family
- DeepSeek
- Parameters
- 3.3B
- Active parameters
- 1.1B per token
- Architecture
- DeepseekOCRForCausalLM
- Context Length
- 8,192 tokens
- Vocabulary Size
- 129,280
- Release Date
- 2025-10-17
- License
- MIT
Frequently Asked Questions
How much VRAM does DeepSeek OCR need?
DeepSeek OCR requires 2.4 GB of VRAM at Q4_K_M, or 7.1 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 3.3B × 4.8 bits ÷ 8 = 2 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 0.8 GB (at full 8K context)
VRAM usage by quantization
What's the best quantization for DeepSeek OCR?
For DeepSeek OCR, Q4_K_M (2.4 GB) offers the best balance of quality and VRAM usage. Q5_0 (2.5 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 1.8 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run DeepSeek OCR on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run DeepSeek OCR. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 2.4 GB of usable unified memory (RAM minus macOS overhead).
Can I run DeepSeek OCR locally?
Yes — DeepSeek OCR can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is DeepSeek OCR?
At Q4_K_M, DeepSeek OCR can reach ~389 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~519 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 0.8 GB active ÷ (8000 × 0.65) = 0.16 ms, plus 12 layers × 0.055 ms = 0.66 ms, so 1000 ÷ 0.82 ms = ~1220 tok/s
Estimated speed at Q4_K_M (2.4 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of DeepSeek OCR?
At Q4_K_M, the download is about 2.00 GB. The full-precision BF16 version is 6.67 GB. The smallest option (Q2_K) is 1.42 GB.
Which GPUs can run DeepSeek OCR?
52 consumer GPUs can run DeepSeek OCR at Q4_K_M. 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 DeepSeek OCR?
163 devices with unified memory can run DeepSeek OCR at Q4_K_M (2.4 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (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.