What are you trying to extract?

Pick your document type. See what actually works.

Invoices & Receipts

Extract line items, totals, vendor info into structured data

Best: Docling (free, local)

Handwritten Notes

Forms, signatures, meeting notes, historical documents

Best: GPT-4o (API)

PDFs & Reports

Multi-page documents, preserve layout, tables, headers

Best: Docling (free, local)

Photos & Screenshots

Camera captures, screen grabs, social media images

Best: PaddleOCR (free, local)

Scanned Books

Digitize printed text, old documents, archives

Best: PaddleOCR-VL (free, local)

ID Cards & Passports

KYC verification, identity documents, MRZ codes

Best: Azure/Google (enterprise)
EXCLUSIVE

We Run Our Own Benchmarks

No vendor claims. Real results. Independently verified.

While others copy numbers from marketing pages, we run the actual benchmarks ourselves. Full datasets. Official evaluation tools. Reproducible results.

1,355 images processed
$2.71 benchmark cost
100% reproducible
Mistral OCR 3 VERIFIED
Composite
79.75
Text
90.1%
Tables
70.9%
Reading Order
91.6%
Dec 19, 2025 View full results →
clearOCR VERIFIED
Composite
31.7
Text
84.6%
Tables
0.8%
Reading Order
86.0%
Dec 19, 2025 Traditional OCR →

Do you have constraints?

Data can't leave my servers
See open source options you can run locally
I need an API I can call
See vendor pricing and accuracy comparison
Speed vs Quality tradeoff
Human preference rankings: ELO scores vs latency

Deep Dives & Techniques

How Docling Works

Understand the architecture of IBM's document understanding library. Why VLM pipelines outperform traditional OCR for complex layouts.

Interactive OCR Correction

A case study on handling OCR "flicker" (H vs N) and camera drift in mobile apps using Google MLKit and centroid anchoring.

Open Source OCR Benchmark

Run on your own servers. No API costs. Full data privacy.

Model OmniDocBench OCRBench (EN) olmOCR License
PaddleOCR-VL
Baidu
 92.86 - 80.0 Apache 2.0
PaddleOCR-VL 0.9B
Baidu
 92.56 - - Apache 2.0
MinerU 2.5
OpenDataLab
 90.67 - 75.2 AGPL-3.0
Qwen3-VL-235B
Alibaba
 89.15 - - Qwen License
MonkeyOCR-pro-3B
Unknown
 88.85 - - Apache 2.0 / MIT
OCRVerse 4B
Unknown
 88.56 - - Apache 2.0 / MIT
dots.ocr 3B
Unknown
 88.41 - 79.1 Apache 2.0 / MIT
Qwen2.5-VL
Alibaba
 87.02 - - Apache 2.0
Chandra v0.1.0
datalab-to
 - - 83.1 Apache 2.0
Infinity-Parser 7B
Unknown
 - - 82.5 Apache 2.0 / MIT
olmOCR v0.4.0
Allen AI
 - - 82.4 Apache 2.0
Marker 1.10.0
VikParuchuri
 - - 76.5 Apache 2.0 / MIT
Marker 1.10.1
VikParuchuri
 - - 76.1 Apache 2.0 / MIT
DeepSeek OCR
DeepSeek
 - - 75.4 Apache 2.0 / MIT
GPT-4o (Anchored)
OpenAI
 - - 69.9 Apache 2.0 / MIT
Nanonets OCR2 3B
Nanonets
 - - 69.5 Apache 2.0 / MIT
Gemini Flash 2
Google
 - - 63.8 Apache 2.0 / MIT
Qwen3-Omni-30B
Alibaba
 - 61.3% - Qwen License
Nemotron Nano V2 VL
NVIDIA
 - 61.2% - NVIDIA Open Model License
CoCa (finetuned)
Google
 - - - Apache 2.0
ViT-G/14
Google
 - - - Apache 2.0
ViT-H/14
Google
 - - - Apache 2.0
ViT-L/16
Google
 - - - Apache 2.0
ViT-B/16
Google
 - - - Apache 2.0
ConvNeXt V2 Huge
Meta
 - - - MIT
ConvNeXt V2 Base
Meta
 - - - MIT
ConvNeXt V2 Tiny
Meta
 - - - MIT
Swin Transformer V2 Large
Microsoft
 - - - MIT
Swin Transformer Large
Microsoft
 - - - MIT
EfficientNetV2-L
Google
 - - - Apache 2.0
EfficientNet-B7
Google
 - - - Apache 2.0
EfficientNet-B0
Google
 - - - Apache 2.0
DeiT-B Distilled
Meta
 - - - Apache 2.0
DeiT-B
Meta
 - - - Apache 2.0
ResNet-152
Microsoft
 - - - MIT
ResNet-50
Microsoft
 - - - MIT
ResNet-50 (A3 training)
Timm
 - - - Apache 2.0
Qwen2.5-VL 72B
Alibaba
 - - - Apache 2.0
CHURRO (3B)
Stanford
 - - - Apache 2.0 / MIT
InternVL2-76B
Shanghai AI Lab
 - - - MIT
InternVL3-78B
Shanghai AI Lab
 - - - Apache 2.0 / MIT
Tesseract
Google (Open Source)
 - - - Apache 2.0
EasyOCR
JaidedAI
 - - - Apache 2.0
Gemini 2.5 Flash
Google
 - - - Apache 2.0 / MIT
olmOCR v0.3.0
Allen AI
 - - - Apache 2.0 / MIT
Qwen2-VL 72B
Alibaba
 - - - Apache 2.0 / MIT
Qwen2.5-VL 32B
Alibaba
 - - - Apache 2.0 / MIT
AIN 7B
Research
 - - - Apache 2.0 / MIT
GPT-4o Mini
OpenAI
 - - - Apache 2.0 / MIT
Azure OCR
Microsoft
 - - - Apache 2.0 / MIT
PaddleOCR
Baidu
 - - - Apache 2.0 / MIT
InternVL3 14B
OpenGVLab
 - - - Apache 2.0 / MIT
o1-preview
OpenAI
 - - - Apache 2.0 / MIT
Llama 3 70B
Meta
 - - - Apache 2.0 / MIT
DeepSeek V3
DeepSeek
 - - - Apache 2.0 / MIT
DeepSeek V2.5
DeepSeek
 - - - Apache 2.0 / MIT
Claude 3.5 Opus
Anthropic
 - - - Apache 2.0 / MIT
AL-Negat
Research
 - - - Apache 2.0 / MIT
GCN
Research
 - - - Apache 2.0 / MIT
Multi-Task Transformer
Research
 - - - Apache 2.0 / MIT
Deep Learning (Heinsfeld)
Research
 - - - Apache 2.0 / MIT
PHGCL-DDGFormer
Research
 - - - Apache 2.0 / MIT
Random Forest
Baseline
 - - - Apache 2.0 / MIT
MAACNN
Research
 - - - Apache 2.0 / MIT
Multi-Atlas DNN
Research
 - - - Apache 2.0 / MIT
Abraham Connectomes
Research
 - - - Apache 2.0 / MIT
Go-Explore
Uber AI
 - - - Apache 2.0 / MIT
BrainGNN
Research
 - - - MIT
MVS-GCN
Research
 - - - Apache 2.0 / MIT
BrainGT
Research
 - - - Apache 2.0 / MIT
SVM with Connectivity Features
Research
 - - - Apache 2.0 / MIT
AE-FCN
Research
 - - - Apache 2.0 / MIT
DeepASD
Research
 - - - Apache 2.0 / MIT
MCBERT
Research
 - - - Apache 2.0 / MIT
ASD-SWNet
Research
 - - - Apache 2.0 / MIT
Agent57
DeepMind
 - - - Apache 2.0 / MIT
MuZero
DeepMind
 - - - Apache 2.0 / MIT
DreamerV3
DeepMind
 - - - Apache 2.0 / MIT
Rainbow DQN
DeepMind
 - - - Apache 2.0 / MIT
DQN (Human-level)
DeepMind
 - - - Apache 2.0 / MIT
Human Professional
Biology
 - - - Apache 2.0 / MIT
BBOS-1
Unknown
 - - - Apache 2.0 / MIT
GDI-H3
Research
 - - - Apache 2.0 / MIT
Plymouth DL Model
Research
 - - - Apache 2.0 / MIT
Co-DETR (Swin-L)
Research
 - - - Apache 2.0 / MIT
InternImage-H
Shanghai AI Lab
 - - - Apache 2.0 / MIT
DINO (Swin-L)
Research
 - - - Apache 2.0 / MIT
YOLOv10-X
Tsinghua
 - - - Apache 2.0 / MIT
Mask2Former (Swin-L)
Meta
 - - - Apache 2.0 / MIT
EfficientDet-D7x
Google
 - - - Apache 2.0 / MIT
CheXNet
Stanford ML Group
 - - - MIT
TorchXRayVision
Cohen Lab
 - - - Apache 2.0
CheXzero
Harvard/MIT
 - - - MIT
MedCLIP
Research
 - - - MIT
GLoRIA
Stanford
 - - - MIT
BioViL
Microsoft
 - - - MIT
RAD-DINO
Microsoft
 - - - MIT
CheXpert AUC Maximizer
Stanford
 - - - Apache 2.0 / MIT
DenseNet-121 (Chest X-ray)
Research
 - - - MIT
ResNet-50 (Chest X-ray)
Research
 - - - MIT
ConVIRT
NYU
 - - - Apache 2.0 / MIT
PatchCore
Amazon
 - - - Apache 2.0
PaDiM
Research
 - - - Apache 2.0
FastFlow
Research
 - - - MIT
EfficientAD
Research
 - - - MIT
SimpleNet
Research
 - - - MIT
DRAEM
Research
 - - - MIT
CFLOW-AD
Research
 - - - Apache 2.0
Reverse Distillation
Research
 - - - MIT
YOLOv8 (Weld Detection)
Ultralytics
 - - - AGPL-3.0
DefectDet (ResNet)
Research
 - - - Apache 2.0 / MIT
When to use open source:
  • - Sensitive data that can't leave your network
  • - High volume processing (no per-page costs)
  • - Offline/air-gapped environments
  • - Full control over the pipeline

Vendor API Benchmark

Pay per page. Fast to integrate. Enterprise support available.

Vendor OmniDocBench OCRBench (EN) olmOCR Price/1k pages
Gemini 2.5 Pro
Google
 88.03 59.3% - varies
Mistral OCR 3
Mistral
 79.75 - 78.0 varies
Mistral OCR 2
Mistral
 - - 72.0 varies
Seed1.6-vision
ByteDance
 - 62.2% - varies
GPT-4o
OpenAI
 - 55.5% - varies
clearOCR
TeamQuest
 31.70 - - varies
Gemini 2.0 Flash
Google
 - - - varies
Gemini 1.5 Pro
Google
 - - - varies
Claude Sonnet 4
Anthropic
 - - - varies
Claude 3.5 Sonnet
Anthropic
 - - - varies
When to use vendor APIs:
  • - Need reasoning/context understanding (GPT-4o, Gemini)
  • - Low volume, occasional use
  • - Need enterprise SLA/support
  • - No infrastructure to maintain

CodeSOTA Score: Cross-Benchmark Comparison

One number to compare models across all benchmarks. Weighted average: primary benchmarks (3x), secondary (2x), language-specific (1x).

How we calculate this

Model Score Cover
OmniDoc
***
OCRBench
***
olmOCR
***
CHURRO
**
CC-OCR
**
KITAB
*
ThaiOCR
*
VideoOCR
*
1 paddleocr-vl
86.4 2/8
93
--
80
--
--
--
--
--
2 dots-ocr-3b
83.8 2/8
88
--
79
--
--
--
--
--
3 mineru-2.5
82.9 2/8
91
--
75
--
--
--
--
--
4 mistral-ocr-3
78.9 2/8
80
--
78
--
--
--
--
--
5 gemini-15-pro
77.1 2/8
--
--
--
--
83
--
--
65
6 gemini-25-pro
72.0 5/8
88
59
--
64
--
--
77
74
7 qwen25-vl-32b
68.8 2/8
--
--
--
--
--
--
77
61
8 qwen25-vl-72b
62.5 3/8
--
--
--
55
--
--
72
69
9 gpt-4o
58.1 5/8
--
56
--
34
76
69
--
66
10 claude-sonnet-4
52.7 2/8
--
--
--
37
--
--
84
--
11 paddleocr-vl-0.9b
-- 1/8
93
--
--
--
--
--
--
--
12 qwen3-vl-235b
-- 1/8
89
--
--
--
--
--
--
--
13 monkeyocr-pro-3b
-- 1/8
89
--
--
--
--
--
--
--
14 qwen25-vl
-- 1/8
87
--
--
--
--
--
--
--
15 ocrverse-4b
-- 1/8
89
--
--
--
--
--
--
--
16 clearocr-teamquest
-- 1/8
32
--
--
--
--
--
--
--
17 seed-1.6-vision
-- 1/8
--
62
--
--
--
--
--
--
18 qwen3-omni-30b
-- 1/8
--
61
--
--
--
--
--
--
19 nemotron-nano-v2-vl
-- 1/8
--
61
--
--
--
--
--
--
20 chandra-ocr-0.1.0
-- 1/8
--
--
83
--
--
--
--
--
21 deepseek-ocr
-- 1/8
--
--
75
--
--
--
--
--
22 marker-1.10.0
-- 1/8
--
--
77
--
--
--
--
--
23 gpt-4o-anchored
-- 1/8
--
--
70
--
--
--
--
--
24 gemini-flash-2
-- 1/8
--
--
64
--
--
--
--
--
25 infinity-parser-7b
-- 1/8
--
--
83
--
--
--
--
--
Open Source
Vendor API
| Scores:
90+
80-89
70-79
60-69
*** Primary | ** Secondary | * Tertiary
Showing top 25 of 44 models with OCR benchmark data 10 models have aggregate scores (2+ benchmarks)

TODO: Priority benchmarks to run

Open source models prioritized (can run locally without API costs):

1. qwen25-vl-32b on omnidocbench (Primary benchmark missing)
2. qwen25-vl-32b on ocrbench-v2 (Primary benchmark missing)
3. qwen25-vl-32b on olmocr-bench (Primary benchmark missing)
4. qwen25-vl-72b on omnidocbench (Primary benchmark missing)
5. qwen25-vl-72b on ocrbench-v2 (Primary benchmark missing)
6. qwen25-vl-72b on olmocr-bench (Primary benchmark missing)
7. internvl3-78b on omnidocbench (Primary benchmark missing)
8. internvl3-78b on ocrbench-v2 (Primary benchmark missing)
9. internvl3-78b on olmocr-bench (Primary benchmark missing)
10. ain-7b on omnidocbench (Primary benchmark missing)

* = open source (run locally) | * = vendor API

OmniDocBench: End-to-end document parsing composite score. OCRBench v2: Overall score across 8 OCR capabilities.

Data from AlphaXiv + Papers With Code.

All Verified Results
258 results with source links

Models

CoCa (finetuned)

OSS

Google

SOTA on ImageNet-1K (91.0%). Combines contrastive and captioning objectives.

Image classificationZero-shot recognition
Contact Google

ViT-G/14

OSS

Google

90.45% top-1 on ImageNet. Giant variant.

High-accuracy classification
Try ViT-G/14 API Contact Google

ViT-H/14

OSS

Google

88.55% top-1 on ImageNet. Huge variant.

High-accuracy classification
Try ViT-H/14 API Contact Google

ViT-L/16

OSS

Google

Large variant. 82.7% with ImageNet-21k pretraining.

Transfer learningFine-tuning
Try ViT-L/16 API Contact Google

ViT-B/16

OSS

Google

Base variant. 81.2% with ImageNet-21k pretraining.

Balanced performanceResearch
Try ViT-B/16 API Contact Google

ConvNeXt V2 Huge

OSS

Meta

88.9% on ImageNet. Best pure ConvNet.

SOTA CNN classification
Try ConvNeXt V2 Huge API Contact Meta

ConvNeXt V2 Base

OSS

Meta

Good balance of speed and accuracy.

Efficient high-accuracy
Try ConvNeXt V2 Base API Contact Meta

ConvNeXt V2 Tiny

OSS

Meta

83.0% on ImageNet. Lightweight variant.

Efficient deployment
Try ConvNeXt V2 Tiny API Contact Meta

Swin Transformer V2 Large

OSS

Microsoft

86.8% on Kinetics-400. Scales to 3B parameters.

High-resolution imagesDense prediction
Try Swin Transformer V2 Large API Contact Microsoft

Swin Transformer Large

OSS

Microsoft

87.3% on ImageNet-1K.

Dense prediction tasks
Try Swin Transformer Large API Contact Microsoft

EfficientNetV2-L

OSS

Google

85.7% on ImageNet. Faster training than V1.

Fast trainingHigh accuracy
Try EfficientNetV2-L API Contact Google

EfficientNet-B7

OSS

Google

84.4% on ImageNet. 8.4x smaller than GPipe.

Best accuracy/params ratio
Try EfficientNet-B7 API Contact Google

EfficientNet-B0

OSS

Google

77.1% on ImageNet. Baseline for compound scaling.

Mobile deploymentEdge devices
Try EfficientNet-B0 API Contact Google

DeiT-B Distilled

OSS

Meta

85.2% on ImageNet. Trained on ImageNet-1K only.

Data-efficient training
Try DeiT-B Distilled API Contact Meta

DeiT-B

OSS

Meta

83.1% on ImageNet without external data.

Training from scratch
Try DeiT-B API Contact Meta

ResNet-152

OSS

Microsoft

78.6% on ImageNet (10-crop). Deep residual network.

Classic baselineTransfer learning
Try ResNet-152 API Contact Microsoft

ResNet-50

OSS

Microsoft

76-80% on ImageNet depending on training. Standard baseline.

Baseline modelBenchmarking
Try ResNet-50 API Contact Microsoft

ResNet-50 (A3 training)

OSS

Timm

80.4% on ImageNet with modern training recipes.

Modern CNN baseline
Try ResNet-50 (A3 training) API Contact Timm

PaddleOCR-VL

OSS

Baidu

#1 on OmniDocBench

Document parsingTablesFormulas
Try PaddleOCR-VL API Contact Baidu

PaddleOCR-VL 0.9B

OSS

Baidu

Lightweight version

Document parsingTables
Try PaddleOCR-VL 0.9B API Contact Baidu

MinerU 2.5

OSS

OpenDataLab

#1 on layout detection (97.5 mAP)

PDF extractionLayout detection
Try MinerU 2.5 API Contact OpenDataLab

Qwen3-VL-235B

OSS

Alibaba

Large model, requires significant compute

Document understandingReasoning
Try Qwen3-VL-235B API Contact Alibaba

MonkeyOCR-pro-3B

OSS

Unknown

Compact model with good performance

Document OCR
Contact Unknown

Gemini 2.5 Pro

API

Google

#1 on OCRBench v2 Chinese, MME-VideoOCR

High accuracyDocument Q&AChinese OCR
Try Gemini 2.5 Pro API Contact Google

Gemini 2.0 Flash

API

Google

#1 on KITAB-Bench (Arabic)

Arabic OCRFast inference
Try Gemini 2.0 Flash API Contact Google

Gemini 1.5 Pro

API

Google

#1 on CC-OCR Multi-Scene

Scene textMultilingualDocument parsing
Try Gemini 1.5 Pro API Contact Google

Qwen2.5-VL

OSS

Alibaba

Document understandingMultilingual
Try Qwen2.5-VL API Contact Alibaba

Qwen2.5-VL 72B

OSS

Alibaba

Document understandingVideo OCRThai OCR
Try Qwen2.5-VL 72B API Contact Alibaba

GPT-4o

API

OpenAI

Best OCR edit distance on OmniDocBench (0.02)

Text extractionDocument Q&A
Try GPT-4o API Contact OpenAI

Seed1.6-vision

API

ByteDance

#1 on OCRBench v2 English

OCR capabilities
Try Seed1.6-vision API Contact ByteDance

Qwen3-Omni-30B

OSS

Alibaba

OCRMultimodal tasks
Contact Alibaba

Nemotron Nano V2 VL

OSS

NVIDIA

Efficient OCREdge deployment
Contact NVIDIA

Chandra v0.1.0

OSS

datalab-to

#1 on olmOCR-Bench (83.1). Best on old scans math, long tiny text, base accuracy.

Document parsingOld scansMath formulas
Try Chandra v0.1.0 API Contact datalab-to

OCRVerse 4B

OSS

Unknown

Strong OmniDocBench performer (88.56)

Document parsingText extraction
Contact Unknown

dots.ocr 3B

OSS

Unknown

Best table TEDS among 3B models. Also #1 on olmOCR tables (88.3)

Document parsingTables
Contact Unknown

Infinity-Parser 7B

OSS

Unknown

PDF parsing
Contact Unknown

olmOCR v0.4.0

OSS

Allen AI

PDF extractionResearch documents
Try olmOCR v0.4.0 API Contact Allen AI

CHURRO (3B)

OSS

Stanford

#1 on CHURRO-DS (82.3 printed, 70.1 handwritten)

Historical documentsHandwritingMultilingual
Try CHURRO (3B) API Contact Stanford

Claude Sonnet 4

API

Anthropic

#1 on ThaiOCRBench

Thai OCRLow hallucination
Try Claude Sonnet 4 API Contact Anthropic

Claude 3.5 Sonnet

API

Anthropic

Lowest hallucination rate on CC-OCR (0.09%)

Document understanding
Try Claude 3.5 Sonnet API Contact Anthropic

InternVL2-76B

OSS

Shanghai AI Lab

Scene textDocument parsing
Contact Shanghai AI Lab

InternVL3-78B

OSS

Shanghai AI Lab

Video OCRDocument understanding
Contact Shanghai AI Lab

Tesseract

OSS

Google (Open Source)

Classic open-source OCR engine

Basic OCROffline use
Try Tesseract API Contact Google (Open Source)

EasyOCR

OSS

JaidedAI

80+ languages supported

MultilingualEasy setup
Try EasyOCR API Contact JaidedAI

DeepSeek OCR

OSS

DeepSeek

DeepSeek's OCR model for document understanding.

Document OCRGeneral OCR
Contact DeepSeek

Marker 1.10.0

OSS

VikParuchuri

Open-source PDF to Markdown converter.

PDF to MarkdownDocument parsing
Try Marker 1.10.0 API Contact VikParuchuri

Marker 1.10.1

OSS

VikParuchuri

Latest version of Marker PDF parser.

PDF to MarkdownDocument parsing
Try Marker 1.10.1 API Contact VikParuchuri

GPT-4o (Anchored)

OSS

OpenAI

GPT-4o with anchored prompting for OCR.

Document understandingOCR
Contact OpenAI

Gemini Flash 2

OSS

Google

Google's fast multimodal model.

Fast inferenceOCR
Contact Google

Gemini 2.5 Flash

OSS

Google

Google's Gemini 2.5 Flash model.

Fast inferenceOCR
Contact Google

olmOCR v0.3.0

OSS

Allen AI

Earlier version of olmOCR.

Document OCRResearch
Try olmOCR v0.3.0 API Contact Allen AI

Mistral OCR 3

API

Mistral

Latest Mistral OCR (Dec 2025). 74% win rate vs OCR 2. Claims 94.9% accuracy. Markdown + HTML table output. $1/1000 pages with batch API.

Document OCRFormsTables
Try Mistral OCR 3 API Contact Mistral

clearOCR

API

TeamQuest

Polish OCR service. Text extraction only - no table/formula recognition. Best for simple documents. VERIFIED by CodeSOTA: 84.6% text accuracy, but 0.8% table TEDS due to lack of structure recognition.

Simple text extractionPolish documentsResearch papers
Try clearOCR API Contact TeamQuest

Mistral OCR 2

API

Mistral

Previous version of Mistral OCR API.

Document OCRFast inference
Try Mistral OCR 2 API Contact Mistral

Nanonets OCR2 3B

OSS

Nanonets

Nanonets' OCR model.

Document OCR
Contact Nanonets

Qwen2-VL 72B

OSS

Alibaba

Qwen2's large vision-language model.

Vision understandingOCR
Try Qwen2-VL 72B API Contact Alibaba

Qwen2.5-VL 32B

OSS

Alibaba

Qwen2.5 32B vision-language model.

Vision understandingOCR
Try Qwen2.5-VL 32B API Contact Alibaba

AIN 7B

OSS

Research

7B parameter OCR model.

OCRDocument understanding
Contact Research

GPT-4o Mini

OSS

OpenAI

Smaller, faster version of GPT-4o.

Cost-effective OCRFast inference
Contact OpenAI

Azure OCR

OSS

Microsoft

Microsoft Azure's OCR service.

Enterprise OCRMulti-language
Contact Microsoft

PaddleOCR

OSS

Baidu

Open-source OCR from PaddlePaddle.

Multilingual OCRChinese text
Try PaddleOCR API Contact Baidu

InternVL3 14B

OSS

OpenGVLab

InternVL3 14B vision-language model.

Vision understandingOCR
Contact OpenGVLab

o1-preview

OSS

OpenAI

OpenAI's reasoning-focused model.

Complex reasoningMath
Contact OpenAI

Llama 3 70B

OSS

Meta

Meta's Llama 3 70B model.

General NLPReasoning
Try Llama 3 70B API Contact Meta

DeepSeek V3

OSS

DeepSeek

DeepSeek's V3 model.

General NLPCode
Contact DeepSeek

DeepSeek V2.5

OSS

DeepSeek

DeepSeek's V2.5 model.

General NLPCode
Contact DeepSeek

Claude 3.5 Opus

OSS

Anthropic

Anthropic's Claude 3.5 Opus model.

Complex reasoningAnalysis
Contact Anthropic

AL-Negat

OSS

Research

Adversarial learning for brain network analysis.

Autism classificationBrain analysis
Contact Research

GCN

OSS

Research

Standard Graph Convolutional Network baseline.

Graph classificationBrain networks
Contact Research

Multi-Task Transformer

OSS

Research

Transformer-based multi-task learning for brain analysis.

Autism classificationMulti-task learning
Contact Research

Deep Learning (Heinsfeld)

OSS

Research

Heinsfeld et al. deep learning approach for ABIDE.

Autism classification
Contact Research

PHGCL-DDGFormer

OSS

Research

Graph transformer with dynamic graph learning.

Autism classificationBrain networks
Contact Research

Random Forest

OSS

Baseline

Standard Random Forest baseline.

Baseline classification
Contact Baseline

MAACNN

OSS

Research

Multi-scale attention CNN for brain imaging.

Autism classification
Contact Research

Multi-Atlas DNN

OSS

Research

DNN combining multiple brain atlases.

Autism classificationMulti-atlas fusion
Contact Research

Abraham Connectomes

OSS

Research

Abraham et al. connectome-based approach.

Brain connectivity analysis
Contact Research

Go-Explore

OSS

Uber AI

Exploration-based reinforcement learning.

Hard exploration gamesAtari
Contact Uber AI

BrainGNN

OSS

Research

ROI-aware graph convolutional layers for interpretable brain network analysis. 73.3% accuracy on ABIDE I.

fMRI analysisBrain connectivityAutism classification
Try BrainGNN API Contact Research

MVS-GCN

OSS

Research

Handles multi-site variability. 69.38% accuracy on ABIDE dataset.

Multi-site brain dataAutism classification
Contact Research

BrainGT

OSS

Research

78.7% AUC on ABIDE dataset, significantly higher than BrainNetTF (73.2%).

Brain disorder diagnosisGraph attention
Contact Research

SVM with Connectivity Features

OSS

Research

70.1% accuracy on ABIDE with functional connectivity features. Classic baseline for brain classification.

Baseline comparisonTraditional ML
Contact Research

AE-FCN

OSS

Research

85% accuracy combining fMRI and sMRI on ABIDE (Rakic et al., 2020).

Feature learningMulti-modal brain data
Contact Research

DeepASD

OSS

Research

93% AUC-ROC on ABIDE-II combining fMRI and SNPs data.

Multi-modal fusionAutism diagnosis
Contact Research

MCBERT

OSS

Research

93.4% accuracy on ABIDE-I with leave-one-site-out cross-validation. Uses phenotypic data.

Medical imagingMulti-modal learning
Contact Research

ASD-SWNet

OSS

Research

76.52% accuracy, 80.65% recall, 0.81 AUC on ABIDE dataset.

Autism diagnosisfMRI classification
Contact Research

Agent57

OSS

DeepMind

First agent to surpass human performance on all 57 Atari games. Uses a meta-controller to adapt exploration.

Hard exploration gamesGeneral arcade gaming
Contact DeepMind

MuZero

OSS

DeepMind

Learns a model of the environment's dynamics without knowing the rules. Mastered Go, Chess, Shogi, and Atari.

Board games (Go, Chess)Atari
Contact DeepMind

DreamerV3

OSS

DeepMind

Scalable world model that masters Atari and Minecraft (MineDojo) with fixed hyperparameters.

Sample efficiencyVisual controlAtari
Contact DeepMind

Rainbow DQN

OSS

DeepMind

Combines 7 improvements to DQN (Double, Dueling, PER, Noisy Nets, Distributional, n-step).

Baseline RLDiscrete control
Contact DeepMind

DQN (Human-level)

OSS

DeepMind

The breakthrough paper (Nature 2015) that started the Deep RL revolution.

Historical baseline
Contact DeepMind

Human Professional

OSS

Biology

Average score of a professional human games tester. Normalized to 100%.

GeneralizationFew-shot learning
Contact Biology

BBOS-1

OSS

Unknown

Achieved massive scores on specific games.

Atari
Contact Unknown

GDI-H3

OSS

Research

Sample efficient benchmark winner.

Atari 100k
Contact Research

Plymouth DL Model

OSS

Research

Up to 98% accuracy on a subset of ABIDE (884 participants). Highlights visual processing regions.

Explainable AIAutism diagnosis
Contact Research

Co-DETR (Swin-L)

OSS

Research

Collaborative Hybrid Assignments Training. SOTA on COCO.

Object Detection
Contact Research

InternImage-H

OSS

Shanghai AI Lab

Large-scale vision model bridging CNN and Transformer.

DetectionSegmentation
Contact Shanghai AI Lab

DINO (Swin-L)

OSS

Research

End-to-end object detection with transformers.

Object Detection
Contact Research

YOLOv10-X

OSS

Tsinghua

NMS-free training for low latency.

Real-time Detection
Contact Tsinghua

Mask2Former (Swin-L)

OSS

Meta

Universal image segmentation architecture.

Segmentation
Contact Meta

EfficientDet-D7x

OSS

Google

Classic efficient detector.

Efficient Detection
Contact Google

CheXNet

OSS

Stanford ML Group

First model to exceed radiologist performance on pneumonia detection. Trained on ChestX-ray14.

Chest X-ray classificationPneumonia detection
Try CheXNet API Contact Stanford ML Group

TorchXRayVision

OSS

Cohen Lab

Pre-trained on 8 datasets (MIMIC, CheXpert, NIH, etc.). Unified 18-pathology output.

Multi-dataset chest X-rayTransfer learning
Try TorchXRayVision API Contact Cohen Lab

CheXzero

OSS

Harvard/MIT

Zero-shot chest X-ray classification using CLIP. No task-specific training needed.

Zero-shot chest X-rayReport generation
Try CheXzero API Contact Harvard/MIT

MedCLIP

OSS

Research

Decoupled contrastive learning on MIMIC-CXR. Semantic matching for medical imaging.

Medical image-text matchingZero-shot diagnosis
Try MedCLIP API Contact Research

GLoRIA

OSS

Stanford

Global-Local Representations for Images using Attention. Learns fine-grained image-text alignment.

Chest X-ray classificationReport generation
Try GLoRIA API Contact Stanford

BioViL

OSS

Microsoft

Biomedical Vision-Language model. Strong performance on phrase grounding.

Medical VQAReport generation
Try BioViL API Contact Microsoft

RAD-DINO

OSS

Microsoft

Self-supervised radiology foundation model. Strong transfer to downstream tasks.

Radiology foundation modelTransfer learning
Contact Microsoft

CheXpert AUC Maximizer

OSS

Stanford

Competition-winning ensemble. 93.0% mean AUC on 5 competition tasks.

CheXpert competitionMulti-label classification
Contact Stanford

DenseNet-121 (Chest X-ray)

OSS

Research

Standard baseline for chest X-ray classification. Pre-trained on ImageNet.

Baseline chest X-rayTransfer learning
Contact Research

ResNet-50 (Chest X-ray)

OSS

Research

Standard ResNet baseline for radiology.

Baseline chest X-ray
Contact Research

ConVIRT

OSS

NYU

Contrastive VIsual Representation learning from Text. Pioneered medical CLIP-like training.

Medical pre-trainingZero-shot transfer
Contact NYU

PatchCore

OSS

Amazon

State-of-the-art on MVTec AD. Uses pretrained features with coreset subsampling.

Industrial inspectionFew-shot anomaly detection
Try PatchCore API Contact Amazon

PaDiM

OSS

Research

Patch-wise anomaly detection using pretrained embeddings and Mahalanobis distance.

Anomaly localizationTexture defects
Try PaDiM API Contact Research

FastFlow

OSS

Research

2D normalizing flows for fast anomaly detection. Good speed-accuracy tradeoff.

Fast inferenceReal-time inspection
Try FastFlow API Contact Research

EfficientAD

OSS

Research

614 FPS inference speed. Optimized for production deployment.

Edge deploymentReal-time industrial
Try EfficientAD API Contact Research

SimpleNet

OSS

Research

Simple yet effective. Competitive with complex methods on MVTec.

Simple deploymentGood generalization
Try SimpleNet API Contact Research

DRAEM

OSS

Research

Discriminatively trained reconstruction for anomaly detection.

Anomaly synthesisPixel-level detection
Try DRAEM API Contact Research

CFLOW-AD

OSS

Research

Real-time unsupervised anomaly detection via conditional normalizing flows.

Precise localizationMulti-scale detection
Try CFLOW-AD API Contact Research

Reverse Distillation

OSS

Research

Reverse distillation for anomaly detection. Strong on texture classes.

Knowledge distillationAnomaly detection
Try Reverse Distillation API Contact Research

YOLOv8 (Weld Detection)

OSS

Ultralytics

Fine-tuned YOLOv8 for weld defect detection. Fast inference for production.

Weld defect detectionReal-time inspection
Try YOLOv8 (Weld Detection) API Contact Ultralytics

DefectDet (ResNet)

OSS

Research

ResNet backbone with FPN for multi-scale defect detection.

Steel defect detectionSurface inspection
Contact Research

Have benchmark results?

Submit your paper or benchmark results. We verify and add them to our database.

Submit Paper

Get OCR updates

New models, benchmark results, and practical guides.

About This Data

All benchmark results are sourced from AlphaXiv benchmark leaderboards. Each data point includes the source URL and access date for verification.

Results marked as "pending verification" are claimed in papers but have not been independently confirmed. We do not include estimated or interpolated values.

JSON API Methodology GitHub