Mistral OCR 3 launches at $2 per 1,000 pages with 74% win rate over previous version
Mistral AI released Mistral OCR 3, a document extraction model priced at $2 per 1,000 pages ($1 with Batch API discount). The model achieves a 74% overall win rate over its predecessor on forms, scanned documents, complex tables, and handwriting according to internal benchmarks.
Mistral OCR 3 launches at $2 per 1,000 pages with 74% win rate over previous version
Mistral AI released Mistral OCR 3 on December 17, 2024, a document extraction model priced at $2 per 1,000 pages, dropping to $1 per 1,000 pages with Batch API discount. The model is now available through API (identifier: mistral-ocr-2512) and via Document AI Playground in Mistral AI Studio.
Performance claims
According to Mistral AI, the model achieves a 74% overall win rate compared to Mistral OCR 2 across forms, scanned documents, complex tables, and handwriting. The company claims state-of-the-art accuracy compared to both enterprise document processing solutions and AI-native OCR solutions, though specific benchmark scores against named competitors were not disclosed.
Mistral evaluated the model using internal benchmarks based on customer use cases, comparing outputs to ground truth using fuzzy-match metrics for accuracy.
Technical capabilities
Mistral OCR 3 extracts text and embedded images from documents, outputting markdown enriched with HTML-based table reconstruction. The model handles:
- Handwriting: Cursive, mixed-content annotations, and handwritten text over printed forms
- Forms: Box detection, labels, handwritten entries, invoices, receipts, compliance forms
- Scanned documents: Compression artifacts, skew, distortion, low DPI, background noise
- Complex tables: Reconstructs structures with headers, merged cells, multi-row blocks, and column hierarchies using HTML table tags with colspan/rowspan
The model supports all languages and document form factors, representing what Mistral describes as a significant upgrade over OCR 2.
Pricing structure
- Standard API: $2 per 1,000 pages
- Batch API: $1 per 1,000 pages (50% discount)
- Annotations: $3 per 1,000 pages
- Self-hosting option available for organizations with data privacy requirements
The model is fully backward compatible with Mistral OCR 2.
What this means
At $2 per 1,000 pages standard pricing, Mistral OCR 3 undercuts typical enterprise document processing solutions that often charge per-page rates in the cents range. The 50% Batch API discount makes it particularly competitive for high-volume workflows. However, without public benchmark comparisons to models from Google, Amazon Textract, or other established OCR providers, customers will need to validate Mistral's performance claims in their own testing. The model's ability to handle multiple document types in a single solution could simplify document processing pipelines that currently require specialized tools for different formats.
Related Articles
Microsoft Releases Mage-VL, a 4B-Parameter Codec-Native Streaming Vision-Language Model
Microsoft has released Mage-VL, a codec-native multimodal foundation model built on a from-scratch 4B-parameter visual encoder paired with Qwen3-4B-Instruct-2507. The model claims up to 3.5x inference speedup over uniform frame sampling and outperforms Qwen3-VL-4B on video and temporal-grounding benchmarks, according to Microsoft.
Unsloth Releases GGUF Quantizations of Kimi K3, a 2.8T-Parameter Open-Weight MoE Model
Unsloth has released GGUF quantizations of Kimi K3, a 2.8-trillion-parameter open-weight Mixture-of-Experts model from Moonshot AI with a 1-million-token context window and native vision support. The largest lossless quantization (Q8) weighs in at 1.56TB.
Anthropic's Claude Opus 5 Generates Full 3D Games From a Single Text Prompt, No Assets Required
Anthropic's Claude Opus 5 can generate playable 3D games, including first-person shooters and Minecraft clones, from a single text prompt with zero external assets. Community tests claim it outperforms GPT-5.6 Sol and Kimi K3 in physics realism and mechanical complexity, though no standardized benchmark has confirmed the comparisons.
ByteDance's Seedance 2.5 Generates 30-Second AI Video Clips With Synced Audio
ByteDance released Seedance 2.5, an AI video model that generates synchronized video and audio in a single pass, producing clips up to 30 seconds long that can be extended further. That's roughly triple the length of Google's Gemini Omni Flash.
Comments
Loading...