GOT OCR2 0 — Hardware Requirements & GPU Compatibility
Vision
GOT-OCR2.0 is a compact 716-million-parameter vision-language model built specifically for OCR, unifying plain-text OCR, formatted-text extraction (tables, formulas, and markup), and fine-grained region- or color-guided recognition in a single end-to-end model rather than a general chat assistant. It reads an image plus a task prompt and outputs recognized text, with an optional rendering mode that reconstructs the original layout as HTML. Unlike two-stage OCR pipelines that pair a separate detector with a text recognizer, GOT-OCR2.0 performs detection-free recognition end to end, and at under a billion parameters it runs comfortably on a single modest consumer GPU or even a CPU. Context length is 32,768 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use. It was published in September 2024 by StepFun alongside its "General OCR Theory" paper.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 32K)Best on 24 GB (Q4_K_M · ~705 tok/s · 32K)Macs from 8 GB (Q4_K_M · ~77 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run GOT OCR2 0
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
vllm serve stepfun-ai/GOT-OCR2_0 --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 Ti (12 GB) (9 GB headroom) · Q4_K_M
- Generation speed
- ~638 tok/s
- generation speed
- Cost per 1M output tokens
- $0.053
- per 1M output tokens
How Much VRAM Does GOT OCR2 0 Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 0.8 GB | 3.8 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 32K context · ~809 tok/s | |
| 0.8 GB | 3.9 GB | 3-bit medium — 3-bit medium quantization | Fits: 32K context · ~771 tok/s | |
| 0.9 GB | 4 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 32K context · ~705 tok/s | |
| 1.0 GB | 4 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 32K context · ~649 tok/s | |
| 1.1 GB | 4.1 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 32K context · ~601 tok/s | |
| 1.2 GB | 4.2 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 32K context · ~537 tok/s | |
| 1.9 GB | 5 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 32K context · ~340 tok/s |
Selected Q4_K_M
4-bit medium quantization2.3GBat 16K context
- Weights
- 0.4 GB
- KV cache
- 1.6 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 21.7 GB free · ~705 tok/s
≥50 tok/s needs ≥71 GB/s · RTX 5090 ~1253, RTX 3090 Ti ~705
Which Hardware Can Run GOT OCR2 0?
Q4_K_M · 0.9 GBGOT OCR2 0 (Q4_K_M) requires 0.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 2+ GB is recommended. Using the full 33K context window can add up to 3.0 GB, bringing total usage to 4.0 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run GOT OCR2 0, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 5.7 GB free · up to 32K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~531Simulate output on RTX 3080RTX 2080 Ti~431Simulate output on RTX 2080 TiRTX 3070 Ti~425Simulate output on RTX 3070 TiM3*8 GB~77Simulate output on iMac 24" M3 (8 GB)Orin Nano~71Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~51Simulate output on iPhone 1720 GB
fits · 17.7 GB free · up to 32K ctx
1 GPU
RX 7900 XT~516Simulate output on RX 7900 XT32–40 GB
fits · 29.7 GB free · up to 32K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~1253Simulate output on RTX 5090M5 Max36 GB~346Simulate output on Mac Studio M5 Max (36 GB)M532 GB~116Simulate output on MacBook Air 13" M5 (32 GB)Orin~143Simulate 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
Similar models
1| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3 0.6BAlibaba | — | 1.7 GB | ~753 | same size |
Specifications
- Publisher
- StepFun
- Parameters
- 716M
- Architecture
- GOTQwenForCausalLM
- Context Length
- 32,768 tokens
- Vocabulary Size
- 151,860
- Release Date
- 2024-09-12
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does GOT OCR2 0 need?
GOT OCR2 0 requires 0.9 GB of VRAM at Q4_K_M, or 1.9 GB at BF16. Full 33K context adds up to 3.0 GB (4.0 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 716M × 4.8 bits ÷ 8 = 0.4 GB
KV Cache + Overhead ≈ 0.5 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 ≈ 3.6 GB (at full 33K context)
VRAM usage by quantization
What's the best quantization for GOT OCR2 0?
For GOT OCR2 0, Q4_K_M (0.9 GB) offers the best balance of quality and VRAM usage. Q5_K_M (1.0 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 0.8 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run GOT OCR2 0 on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run GOT OCR2 0. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 0.9 GB of usable unified memory (RAM minus macOS overhead).
Can I run GOT OCR2 0 locally?
Yes — GOT OCR2 0 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is GOT OCR2 0?
At Q4_K_M, GOT OCR2 0 can reach ~5161 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~705 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 ÷ 0.93 × 0.65 = ~5591 tok/s
Estimated speed at Q4_K_M (0.9 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of GOT OCR2 0?
At Q4_K_M, the download is about 0.43 GB. The full-precision BF16 version is 1.43 GB. The smallest option (Q2_K) is 0.30 GB.
Which GPUs can run GOT OCR2 0?
52 consumer GPUs can run GOT OCR2 0 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 GOT OCR2 0?
163 devices with unified memory can run GOT OCR2 0 at Q4_K_M (0.9 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.