Model support matrix
Large Language Models (LLMs) integration matrices define the validation depth, testing scopes, and cloud provider compatibilities for foundational models verified within Agent Studio.
To ensure coverage and maintain confidence in model behavior, the support matrix classifies
validated foundational models from AWS Bedrock, Azure OpenAI, and Cloudera AI Inference service
into the following verification tiers:
- End-to-End (E2E) — Covers full lifecycle testing of complex application workflows, rigorous guardrail validation, latency thresholds, and structural output consistency. Cloudera recommends this tier for production-critical environments.
- Smoke Testing — Covers basic connectivity and readiness validation, confirming API endpoint responsiveness, core schema adherence, and foundational token processing capabilities under non-complex workloads.
Review the following model support matrix requirements before configuring production-critical workflows:
| Provider | Model identifier or Endpoint | Testing scope | Access key tier |
|---|---|---|---|
| Cloudera AI Inference | nvidia/nemotron-3-super-120b-a12b |
E2E | Official |
| Cloudera AI Inference | nvidia/llama-3.3-nemotron-super-49b-v1.5 |
Smoke | Official |
| AWS Bedrock | global.anthropic.claude-haiku-4-5-20251001-v1:0 |
Smoke | Official |
| AWS Bedrock | global.anthropic.claude-opus-4-6-v1 |
E2E | Official |
| AWS Bedrock | global.anthropic.claude-opus-4-7 |
Smoke | Official |
| AWS Bedrock | global.anthropic.claude-opus-4-8 |
Smoke | Official |
| AWS Bedrock | global.anthropic.claude-sonnet-4-6 |
Smoke | Official |
| Azure Open AI | gpt-4o-mini |
E2E | Official |
| Azure Open AI | gpt-4o |
E2E | Official |
| Azure Open AI | o1 |
Smoke | Official |
| OpenAI Compatible | vertex_ai/gemini-3-flash-preview |
Smoke | Official |
| OpenAI Compatible | azure/gpt-5.2 |
Smoke | Official |
| OpenAI Compatible | azure/cloudops-gpt-5.4 |
Smoke | Official |
