OvisOCR2 — Hardware Requirements & GPU Compatibility
Vision
OvisOCR2 is a compact 853-million-parameter end-to-end vision-language model for page-level document OCR, built by post-training Qwen3.5-0.8B with a data engine that mixes real and synthetic documents and a multi-stage supervised fine-tuning, reinforcement learning, and OPD training recipe. Given a document page image it outputs a single Markdown document that reproduces the natural reading order, including body text, formulas rendered as LaTeX, tables as HTML, and captioned image regions, replacing traditional multi-stage OCR pipelines with one model. It set a new state of the art on the OmniDocBench v1.6 leaderboard, the first end-to-end model to top a benchmark long dominated by pipeline-based approaches, and also leads PureDocBench. At under a billion parameters it runs comfortably on a single consumer GPU or even a CPU. Context length is 262,144 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in July 2026.
217.4K downloads
470 likesSpecifications
- Publisher
- ATH-MaaS
- Parameters
- 853M
- Architecture
- Qwen3_5ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-07-13
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does OvisOCR2 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 0.7 GB | 7.1 GB | 0.36 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 0.7 GB | 7.1 GB | 0.37 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 0.8 GB | 7.2 GB | 0.42 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 0.8 GB | 7.2 GB | 0.43 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 0.9 GB | 7.3 GB | 0.51 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 1.0 GB | 7.3 GB | 0.61 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 1.1 GB | 7.5 GB | 0.70 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 1.2 GB | 7.6 GB | 0.85 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run OvisOCR2?
Q4_K_M · 0.9 GBOvisOCR2 (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 262K context window can add up to 6.4 GB, bringing total usage to 7.3 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run OvisOCR2?
Q4_K_M · 0.9 GB59 devices with unified memory can run OvisOCR2, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomWhere to Download OvisOCR2
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 OvisOCR2 need?
OvisOCR2 requires 0.9 GB of VRAM at Q4_K_M, or 2.1 GB at BF16. Full 262K context adds up to 6.4 GB (7.3 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 853M × 4.8 bits ÷ 8 = 0.5 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 6.8 GB (at full 262K context)
VRAM usage by quantization
Q4_K_M0.9 GBQ4_K_M + full context7.3 GB- What's the best quantization for OvisOCR2?
For OvisOCR2, Q4_K_M (0.9 GB) offers the best balance of quality and VRAM usage. Q4_K_L (0.9 GB) provides better quality if you have the VRAM. The smallest option is IQ2_M at 0.6 GB.
VRAM requirement by quantization
IQ2_M0.6 GBIQ3_M0.7 GBIQ4_NL0.8 GBQ4_K_M ★0.9 GBQ5_K_S0.9 GBBF162.1 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run OvisOCR2 on a Mac?
OvisOCR2 requires at least 0.6 GB at IQ2_M, 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 OvisOCR2 locally?
Yes — OvisOCR2 can run locally on consumer hardware. At Q4_K_M quantization it needs 0.9 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is OvisOCR2?
At Q4_K_M, OvisOCR2 can reach ~5581 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~762 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 ÷ 0.9 × 0.65 = ~6047 tok/s
Estimated speed at Q4_K_M (0.9 GB)
~6047 tok/s~762 tok/s~6047 tok/s~5581 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of OvisOCR2?
At Q4_K_M, the download is about 0.51 GB. The full-precision BF16 version is 1.71 GB. The smallest option (IQ2_M) is 0.29 GB.
- Which GPUs can run OvisOCR2?
52 consumer GPUs can run OvisOCR2 at Q4_K_M (0.9 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 OvisOCR2?
59 devices with unified memory can run OvisOCR2 at Q4_K_M (0.9 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.