AI research resources

CARC does not currently provide hosted AI, large language model (LLM), or agent services as a production service. Researchers whose projects require AI models, API access, or agent-oriented workflows can request or explore research-focused services through the external programs below.

These resources may be useful for faculty principal investigators, postdoctoral researchers, graduate students, undergraduate students working on mentored research, and research staff. Each service has its own eligibility rules, allocation process, acceptable-use requirements, and account-review workflow.

External resources

NAIRR Pilot

The NAIRR Pilot allocation opportunities page is a starting point for researchers seeking access to national AI research resources. It is intended for research and education projects that need AI-related computing, model, data, or platform access beyond what a local department or campus service can provide.

ACCESS

ACCESS account registration is the entry point for researchers who want to use NSF-funded advanced cyberinfrastructure. ACCESS can be appropriate when a project needs national research computing resources, allocations, or services that support larger computational workflows.

AI-VERDE

AI-VERDE provides documentation for using CyVerse’s AI-VERDE services, including an OpenAI-compatible API for connecting tools and code to available models. Researchers who need API-based access for scripts, notebooks, or agent workflows may find this a useful place to start.

Who can use these resources

Access usually depends on the service and the research context. In general, these programs are aimed at academic research and education users such as PIs, postdocs, graduate students, students working under faculty supervision, and research staff. Some requests may need a project description, sponsoring PI, institutional affiliation, or explanation of the computational and AI requirements.

Before applying, be ready to describe:

  • the research goal or education use case;
  • why AI, LLM, API, or agent access is needed;
  • expected data sensitivity and privacy considerations;
  • the scale of the requested work, such as model access, computing time, or workflow requirements;
  • the people who need access and their roles on the project.

Get help

The CARC Research Facilitation team is happy to talk through options, help scope the technical requirements, and point researchers toward the most appropriate external service. For questions, email researchtech@unm.edu or visit office and consultation hours.