The AI Agent Index

Documenting the technical and safety features of deployed agentic AI systems

h2oGPTe


Basic information

Website: https://web.archive.org/web/20241227125638/https://h2o.ai/platform/enterprise-h2ogpte/

Short description: The industry’s first multi-agent Generative AI platform to bring together the strengths of Generative AI and Predictive AI with airgapped, on-premise deployment options [source]

Intended uses: What does the developer say it’s for? Generative AI platform for businesses [source]

Date(s) deployed: Multiple versions exist, some date back to Dec 12 2023 [source]


Developer

Website: https://web.archive.org/web/20250102014554/https://h2o.ai/

Legal name: H2O.AI, Inc [source]

Entity type: Corporation [source]

Country (location of developer or first author’s first affiliation): Incorporation: Delaware, USA (H2O.AI, INC. 5060671) [source]

Safety policies: What safety and/or responsibility policies are in place? Unknown


System components

Backend model: What model(s) are used to power the system? The primary model is Claude 3.5 Sonnet, but it can use other models [source]

Publicly available model specification: Is there formal documentation on the system’s intended uses and how it is designed to behave in them? None

Reasoning, planning, and memory implementation: How does the system ‘think’? Next steps are determined by “intelligent model routing” [source]. A variety of architectural modules are used [source].

Observation space: What is the system able to observe while ‘thinking’? Text, audio, and images [source]

Action space/tools: What direct actions can the system take? Depending on the use case, it can write code, create files, and chat with the user [source]

User interface: How do users interact with the system? Primarily a chat interface [source]

Development cost and compute: What is known about the development costs? Unknown


Guardrails and oversight

Accessibility of components:

  • Weights: Are model parameters available? Unknown
  • Data: Is data available? Unknown
  • Code: Is code available? Closed source
  • Scaffolding: Is system scaffolding available? Partially available [source]
  • Documentation: Is documentation available? Available [source]

Controls and guardrails: What notable methods are used to protect against harmful actions? Supports customizable guardrails against prompts, types of responses, regex patterns, presidio labels, PII labels [source]

Customer and usage restrictions: Are there know-your-customer measures or other restrictions on customers? None

Monitoring and shutdown procedures: Are there any notable methods or protocols that allow for the system to be shut down if it is observed to behave harmfully? Unknown


Evaluation

Notable benchmark evaluations: 65% on GAIA [source]

Bespoke testing: Guide [source]

Safety: Have safety evaluations been conducted by the developers? What were the results? None

Publicly reported external red-teaming or comparable auditing:

  • Personnel: Who were the red-teamers/auditors? None
  • Scope, scale, access, and methods: What access did red-teamers/auditors have and what actions did they take? None
  • Findings: What did the red-teamers/auditors conclude? None

Ecosystem information

Interoperability with other systems: What tools or integrations are available? Integrates with H2O feature store [source]

Usage statistics and patterns: Are there any notable observations about usage? Customer case studies available for the company [source]


Additional notes

None