The OCR Wars Heat Up: Baidu Releases PaddleOCR-VL-1.5 to Challenge DeepSeek
Just days after DeepSeek-OCR 2's release, Baidu's PaddleOCR team strikes back with PaddleOCR-VL-1.5, claiming superior performance in document parsing.
The OCR Wars Heat Up: Baidu Releases PaddleOCR-VL-1.5
Jan 30, 2026
The battle for visual language model (VLM) dominance isn't stopping at LLMs. Just days after DeepSeek released its acclaimed DeepSeek-OCR 2, tech giant Baidu has responded with a major update to its open-source toolkit: PaddleOCR-VL-1.5.
What's New?
Released quietly on GitHub on January 29, 2026, this new version targets the exact same niche as DeepSeek's latest offering: high-precision document parsing and structure extraction.
Initial benchmarks released by the Paddle team suggest that PaddleOCR-VL-1.5 may edge out DeepSeek-OCR 2 in specific tasks:
- Table Extraction: Claimed 5% higher accuracy on complex financial tables.
- Efficiency: Optimized for edge deployment, running faster on consumer-grade GPUs.
- Multilingual Support: Expanded support for mixed-language documents.
DeepSeek OCR 2 vs. PaddleOCR-VL-1.5
| Feature | DeepSeek OCR 2 | PaddleOCR-VL-1.5 |
|---|---|---|
| Release Date | Jan 27, 2026 | Jan 29, 2026 |
| Focus | Visual CoT & Reasoning | Structure Parsing & Speed |
| Architecture | DeepEncoder V2 | Modified NaViT |
| Open Source | MIT License | Apache 2.0 |
Community Reaction
The timing is undeniable. "It's clearly a response," says one developer on Hacker News. "DeepSeek set a new bar on Tuesday, and Baidu tried to clear it on Thursday."
We are currently running our own internal benchmarks to verify these claims. While DeepSeek-OCR 2 focuses heavily on the "reasoning" aspect of reading (understanding what it reads), Baidu seems to be doubling down on the "structural" accuracy (getting the layout perfect).
Stay tuned for our full comparison review next week.
Follow DeepSeek V4 App for the latest AI model news.
Author
More Posts
OpenAI GPT-5.4 Drops: 1M Context + Native Agents to Block DeepSeek V4!
OpenAI launched its flagship GPT-5.4 with 1 million native context and an agentic engine, aiming to build a technical moat before the DeepSeek V4 release.
The Hardcore Truth Behind DeepSeek V4's Delayed Release
Why did DeepSeek V4 miss its March 2nd launch window? Exploring the truth behind the delay: domestic compute migration, multimodal integration, and strategic timing.
Battle of Lightweight Models: GPT-5.3 Instant and Gemini 3.1 Flash-Lite Arrive—How Can DeepSeek V4 Stay Ahead?
With OpenAI and Google releasing GPT-5.3 Instant and Gemini 3.1 Flash-Lite on the same day, the lightweight model market is boiling over. This article analyzes the impact of these models on Agent ecosystems like OpenClaw and DeepSeek V4's core competitive advantages in this changing landscape.
Newsletter
Join the community
Subscribe to our newsletter for the latest news and updates