AI Paper Atlas

Research map / Benchmarks / OCRBench

OCRBench leaderboard

58 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on OCRBench. Best score: 907 (Full CE-OCR Framework, CVPR 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

OCRBench · score

45 papers, each with its best reported number for its own method; higher is better.

#score ↑MethodPaperTopic
1907Full CE-OCR FrameworkConsensus Entropy: Harnessing Multi-VLM Agreement for Self-Verifying and Self-Improving OCR
CVPR 2026 · Table 6
Multimodal LLMs
2884Game-R1 7B
Qwen2.5-VL-7B
Game Ground Bench: Probing the Limits of LVLMs in Complex Semantic Grounding Across Game Universes
AAAI 2026 · Table 3
Multimodal LLMs
3882DeepEyesV2
Qwen2.5-VL-7B
DeepEyesV2: Toward Agentic Multimodal Model
ICLR 2026 · Table 2
Multimodal LLMs
4871DeFacto
Qwen2.5-VL-7B
DeFacto: Counterfactual Thinking with Images for Enforcing Evidence-Grounded and Faithful Reasoning
ICML 2026 · Table 3
Multimodal LLMs
5869Eagle2.5-8B
Qwen2.5-7B
Eagle 2.5: Boosting Long-Context Post-Training for Frontier Vision-Language Models
NeurIPS 2025 · Image benchmarks table
Multimodal LLMs
6866Qwen2.5-VL-7B, increased perception iterations
Qwen2.5-VL-7B
VideoChat-R1.5: Visual Test-Time Scaling to Reinforce Multimodal Reasoning by Iterative Perception
NeurIPS 2025 · text
Multimodal LLMs
7855AdaPatch
SAIL-VL-2B
One Patch Doesn’t Fit All: Adaptive Patching for Native-Resolution Multimodal Large Language Models
ICLR 2026 · Table 1
Multimodal LLMs
8849ORR (MMAT-1M)
InternVL2.5-8B
MMAT-1M: A Large Reasoning Dataset for Multimodal Agent Tuning
ICCV 2025 · Table 3
Multimodal LLMs
9833UniARUnified Multimodal Autoregressive Modeling with Shared Context—Visual Tokenizer is Key to Unification
ICML 2026 · Table 4
Human and embodied modeling
10829BlueLM-V-3B
BlueLM-3B
BlueLM-V-3B: Algorithm and System Co-Design for Multimodal Large Language Models on Mobile Devices
CVPR 2025 · Table 3
Quantization
35 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

OCRBench · accuracy

13 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
190.7Qwen2.5-VL-7B + ECRD
Qwen2.5-VL-7B
See It, Say It, Sorted: An Iterative Training-Free Framework for Visually-Grounded Multimodal Reasoning in LVLMs
CVPR 2026 · Table 5
Multimodal LLMs
289.8AdaThinking-E (ATK)
Qwen3-VL-8B
AdaThinking-E: One-Token Entropy Regulation for Adaptive Thinking
ECCV 2026 · Table 1
Reasoning and chain of thought
387.6FactoSR-8B-RL
Qwen3-VL-8B-Instruct
Unfold The World: Factorize 4D Properties in Reinforcing Spatial Understanding
ECCV 2026 · Table 2
Multimodal LLMs
486.3Manzano-30BMANZANO: A Simple and Scalable Unified Multimodal Model with a Hybrid Vision Tokenizer
ICLR 2026 · Table 2
Multimodal LLMs
586.3Thyme-VL-7B
Qwen2.5-VL-7B
Thyme: Think Beyond Images
ICLR 2026 · Table 1
Multimodal LLMs
685.3SeProD (Mini-o3 search, 3B prophet)
Mini-o3
Self-Prophetic Decoding to Unlock Visual Search in LVLMs
ICML 2026 · Table 2
Multimodal LLMs
783.6R-4B-RL (A-T)
R-4B
R-4B: Incentivizing General-Purpose Auto-Thinking in MLLMs via Bi-Mode Annealing and Reinforce Learning
CVPR 2026 · Table 1
Reasoning and chain of thought
879.7Tuna-2
Qwen2.5-7B-Instruct
Tuna-2: Pixel Embeddings Beat Vision Encoders for Multimodal Understanding and Generation
NeurIPS 2026 · Table 1
Human and embodied modeling
969.9FLARE-X 8B
Llama-3.1-8B
FLARE: Fully Integration of Vision-Language Representations for Deep Cross-Modal Understanding
ICLR 2026 · Table 1
Cross-modal learning
1066.3UIX-Qwen2
Qwen2-7B-Instruct
Harnessing Webpage UIs for Text-Rich Visual Understanding
ICLR 2025 · Table: Results on general OCR/
Multimodal LLMs
3 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.