Austrian Academy of Sciences - Mistral AI

2026-05-25 · Source: mistral.ai via Google News · Field: Science & Research — Social Sciences & Behavioral Studies, Research Methodology & Innovation · Depth: Fundamental Awareness, short

Summary

The Austrian Academy of Sciences (OeAW), in partnership with Mistral AI and Reply, has developed Apollo, the world's first advanced Large Language Model specifically for Ancient Greek. Launched on May 25, 2026, Apollo is trained on a specialized corpus. This corpus comprises 600 million words from historical Greek texts, alongside tens of thousands of published inscriptions and papyri. This AI system aims to revolutionize the study of antiquity by enabling researchers to reconstruct over one million unread Greek papyri. It accelerates discoveries from years to hours, with future phases planned for semantic search and handwritten inscription decipherment. This collaboration provides a secure, European-based infrastructure, allowing scholars to overcome manual transcription limitations and unlock knowledge trapped in ancient fragments.

Key takeaway

For Research Scientists in humanities grappling with vast, fragmented historical texts, Apollo demonstrates how specialized AI can transform your workflow. You should consider custom LLMs trained on domain-specific corpora to accelerate decipherment and analysis from years to hours. This approach frees you to focus on interpretation, opening new avenues for understanding complex historical data and preserving cultural heritage.

Key insights

Apollo, an LLM trained on 600M Greek words, revolutionizes ancient text decipherment, accelerating research from years to hours.

Principles

• AI can amplify understanding of history.
• Specialized LLMs unlock domain-specific knowledge.
• Secure infrastructure is vital for sensitive data.

Method

The OeAW partnered with Mistral and Reply to train Apollo, a custom LLM, on a 600 million-word corpus of historical Greek texts, inscriptions, and papyri, ensuring a secure, European-based infrastructure.

In practice

• Reconstruct damaged ancient texts.
• Identify thematic connections across collections.
• Conduct semantic searches on historical data.

Topics

• Large Language Models
• Ancient Greek
• Digital Humanities
• Text Reconstruction
• Papyrology
• Mistral AI

Best for: AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• Towards Generalist Agents for Accelerating Scientific Discovery171 — LLMs can accelerate scientific discovery by generating and refining hypotheses within agentic, iterative workflows.
• Custom model training. Domain-specific language models. | Mistral - Mistral AI — Mistral AI enables enterprises to build fully-owned, domain-specialized LLMs through collaborative customization and flexible deployment.
• PRL-Bench: A Comprehensive Benchmark Evaluating LLMs' Capabilities in Frontier Physics Research — LLMs struggle with autonomous, long-horizon physics research, primarily due to domain knowledge gaps and reasoning instability.
• Mar 23, 2026ScienceVibe physics: The AI grad student — AI can significantly accelerate expert-guided theoretical physics research, but requires rigorous human oversight.
• Opportunities and Challenges of Large Language Models for Low-Resource Languages in Humanities Research — LLMs offer transformative potential for low-resource language research in humanities, despite significant data and ethical challenges.
• Google I/O showed how the path for AI-driven science is shifting — Google is shifting its AI for science focus from specialized tools to agentic, LLM-based systems for autonomous research.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by mistral.ai via Google News.