Supported models and providers.

Stackmint is model-agnostic. Agents run across foundation models, inference platforms, cloud runtimes and developer tools, so each workflow can use the right stack for performance, cost, compliance or on-premise constraints.

Foundation models

12
  • OpenAIFoundation model provider
  • AnthropicFoundation model provider
  • ClaudeFoundation model provider
  • Mistral AIFoundation model provider
  • DeepSeekFoundation model provider
  • QwenFoundation model provider
  • ZhipuFoundation model provider
  • Gemma (Google)Foundation model provider
  • GrokFoundation model provider
  • GroqFoundation model provider
  • PerplexityFoundation model provider
  • Meta LlamaFoundation model provider

Multimodal and creative models

8
  • MidjourneyMultimodal, vision, creative
  • StabilityMultimodal, vision, creative
  • RunwayMultimodal, vision, creative
  • LumaMultimodal, vision, creative
  • Dream MachineMultimodal, vision, creative
  • SunoMultimodal, vision, creative
  • IdeogramMultimodal, vision, creative
  • SoraMultimodal, vision, creative

Inference and model hubs

4
  • Hugging FaceInference and model hub
  • Hugging Face InferenceInference and model hub
  • ReplicateInference and model hub
  • CohereInference and model hub

Cloud and infrastructure providers

13
  • AWSCloud, infrastructure
  • AWS BedrockCloud, infrastructure
  • Azure AICloud, infrastructure
  • MicrosoftCloud, infrastructure
  • Google CloudCloud, infrastructure
  • Google GeminiCloud, infrastructure
  • Google Vertex AICloud, infrastructure
  • NVIDIACloud, infrastructure
  • CloudflareCloud, infrastructure
  • Cloudflare Workers AICloud, infrastructure
  • SnowflakeCloud, infrastructure
  • VercelCloud, infrastructure
  • v0.devCloud, infrastructure

Developer and workflow tools

5
  • LangChainDeveloper tools, orchestration
  • LangGraphDeveloper tools, orchestration
  • LangfuseDeveloper tools, orchestration
  • LlamaIndexDeveloper tools, orchestration
  • MCPDeveloper tools, orchestration

Why multi-model support matters

Different workflows need different models. Some require top-tier reasoning, others need real-time latency, GPU-optimized inference, multimodal capabilities or regional deployments. Stackmint lets you route agents to the best provider without rewriting Buds or Branches.

You can enable or disable providers per tenant or workspace. Stackmint handles orchestration, observability, retries and consistency, while you stay in full control of your model and infrastructure choices.

See the platform

Need a provider we don’t list yet?

We add new providers based on customer demand. If you’d like Stackmint to support another model, inference endpoint or self-hosted runtime (including vLLM, TGI, SGLang, Ollama or custom deployments), reach out.

Request a provider: builders@stackmint.ai