We are excited to announce the release of h2oGPTe 1.6. Read on to learn about the new features and improvements in h2oGPTe 1.6.x series that help you find answers and generate content based on your private data.
This patch release fixes a crash loop in FIPS-enabled deployments and a PDF parsing crash in FedRAMP and offline environments.
This patch release adds Gemini 3.1 Pro Preview and Claude 4.6 models, an optimized collection import option, and security updates.
This patch release adds chat generation control endpoints, faster document counting, and permission and citation improvements.
This patch release adds inline reference pills with PDF highlighting, RAG query filters, Confluence attachments, and GPT-5.2 and Gemini-3-pro-preview support.
This patch release adds a math formula display toggle, improves chat history display, and fixes text-to-speech playback.
This patch release improves markdown math rendering, error message clarity, and workspace validation.
This patch release moves the chat toolbar into the input area, redesigns connector configuration, and fixes SSO username display and workspace assignment issues.
This patch release adds workspaces, the Tool Builder Agent, video scene descriptions, auto-sync collection connectors, and new model support.
This patch release adds Spanish language support, a Data Science Agent type, a centralized Secret Manager, and document processing performance improvements.
This patch release adds prompt template group sharing and configurable banners, and improves permissions, guest authentication, and concurrent processing.
This patch release improves the mobile experience and API documentation, and adds support for new models including Claude 3.5 (Bedrock), Gemini 2.0 Flash, and DeepSeek V3 and R1.
This patch release adds Lite ingest mode, speeds up PDF ingestion and chat queries, and fixes agent chat session issues.
The 1.6.0 release introduces the h2oGPTe Agent, role-based access control, an OpenAPI-compliant REST API, and Custom GPT creation.