(BETA Environment)
| Module |
Provider |
Proxy API |
Model |
Chat API |
Assistant Module |
RAG Module |
Frontend |
| saia.models.openai |
OpenAI |
✅ |
* (1) |
❌ |
✅ |
✅ |
✅ |
| saia.models.googlegenai |
Google Generative AI |
✅ |
gemini-1.0-pro |
❌ |
✅ |
✅(2) |
❌ |
| saia.models.googlevertexai |
Google VertexAI |
✅ |
gemini-1.0-pro |
✅ |
✅ |
✅(3) |
✅ |
|
gemini-1.5-pro-preview-0409
|
✅ |
✅ |
✅ |
| pending |
Anthropic |
✅ |
claude-3-opus-20240229 |
✅ |
✅ |
✅(4) |
✅ |
|
claude-3-sonnet-20240229
|
✅ |
✅ |
✅ |
| |
|
|
|
|
|
|
|
| pending |
Azure OpenAI |
✅ |
* (1) |
❌ |
✅ |
✅(5) |
✅ |
| saia.models.nvidia-nemo |
NVidia Nemo |
✅ |
NVidia Models (2) |
|
|
|
|
(1) - Needs to be enabled by request.
(2) - Set the RAG LLM Provider as google.
(3) - Set the RAG LLM Provider as googlevertexai.
(4) - Set the RAG LLM Provider as anthropic.
(5) - Set the RAG LLM Provider as azureopenai.
| Model |
Proxy API |
Chat API |
Assistant Module |
RAG Module(2) |
Frontend |
| code-llama-70B |
✅ |
❌ |
✅ |
✅ |
✅ |
| code-llama-13B |
| code-llama-34B |
| NV-Llama2-70B-RLHF |
❌ |
| NV-Llama2-70B-SteerLM-Chat |
✅ |
| mamba-chat |
✅ |
| Nemotron-3-8B-QA |
❌(1) |
| Nemotron-3-8B-Chat-SteerLM |
✅ |
| Mixtral-8x7B-Instruct |
✅ |
| gemma-2B |
✅ |
| gemma-7B |
✅ |
| Yi-34B |
✅ |
| NeVA-22B |
❌(1) |
| Fuyu-8B |
❌(1) |
| Llama-2-70B |
✅ |
| Llama-2-13B |
| meta/llama3-8b-instruct |
(4) |
|
✅(3) |
✅(3) |
✅(3) |
| meta/llama3-70b-instruct |
|
| snowflake/arctic |
|
✅(3) |
| databricks/dbrx-instruct |
|
✅(3) |
| mistralai/mistral-7b-instruct-v0.2 |
|
✅(3) |
| mistralai/mistral-large |
|
| mistralai/mixtral-8x22b-instruct-v0.1 |
|
| microsoft/phi-3-mini-128k-instruct |
|
✅(3) |
(1) - Not a chat model
(2) - The provider to set for the RAG LLM Settings section must be: nvidia.nemo and the Model Name must match one of the supported models detailed.
(3) - Available since Nvidia module Revision #4.
(4) - Evaluation pending