Document Processing · Enterprise Knowledge Search
Researchers at the Austrian Academy of Sciences needed a way to analyze and reconstruct huge volumes of fragmented Ancient Greek texts that manual transcription and interpretation could not process quickly or at sufficient scale.
Reported impact
Apollo lets researchers reconstruct damaged Ancient Greek texts in hours instead of years, opens a path to working through more than one million unread papyri, and creates a foundation for semantic search and future handwritten inscription decipherment.
The Austrian Academy of Sciences is building Apollo, a custom large language model for Ancient Greek, to turn large-scale classical text analysis into an interactive AI workflow. The project combines a 600 million-word historical Greek corpus with tens of thousands of published inscriptions and papyri so the model can learn the linguistic variation needed to reconstruct missing passages and surface thematic links across collections. Mistral contributed model-building expertise and, with services partner Reply, helped establish a secure European-based infrastructure for training and use. Researchers can now apply Apollo to damaged texts that previously required years of manual transcription and interpretation, and the platform is being extended toward semantic search and handwritten inscription decipherment so scholars can move from recovery of fragments to broader historical analysis.
Historical Greek texts, published inscriptions, and papyri feed Apollo's training corpus; researchers submit fragmentary or damaged text to the model, Apollo proposes restorations and semantic matches, and the outputs let scholars move into interpretation and broader historical analysis faster.
"Apollo, the world's first advanced Large Language Model for Ancient Greek."
"The collaboration ensured a secure, European-based infrastructure"
"Trained on the largest digital corpus of historical Greek to date"
"What once required years of painstaking work can now be achieved in hours"
"with future phases planned to support semantic search and handwritten inscription decipherment"
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
OpenAI Sample Aug 31, 2026 Polimill Polimill needed to reduce administrative burden across fragmented Japanese municipal systems so public employees could find prior decisions, prepare assembly responses, and deliver services without each municipality… 3-5x development speed increase from Codex and OpenAI hands-on support Codex · GPT models · OpenAI API 
NVIDIA Sample Jul 7, 2026 Public Ministry of Rio Grande do Sul (MPRS) MPRS needed to manage rising legal caseloads, archived police inquiries, and public-service requests more efficiently, but the volume of legal documents and the need for compliant, locally controlled analysis made… 497 municipalities Archiving Intelligence Assistant (AIA) · Citizen Access Agent · Justice Intelligence platform 
Mistral Sample Apr 24, 2026 Ardian Ardian needed to analyze large volumes of confidential due diligence material faster, but its teams were overwhelmed by scattered internal documents, NDA constraints, and traditional machine-learning approaches that did… 18,000+ documents since inception GAIA · Mistral models 