Model Coverage

Model Coverage Overview

View as Markdown

NeMo AutoModel integrates with Hugging Face transformers. Any LLM or VLM that can be instantiated through transformers can also be used using NeMo AutoModel, subject to runtime, third-party software dependencies, and feature compatibility.

Supported Hugging Face Auto Classes

Auto ClassTaskStatusDetails
AutoModelForCausalLMText Generation (LLM)SupportedSee LLM model list.
Block-Diffusion LLMsText Generation (Diffusion LLM)SupportedSee Diffusion LLM model list.
AutoModelForImageTextToTextImage-Text-to-Text (VLM)SupportedSee VLM model list.
Custom multimodal modelsUnified multimodal trainingSupportedSee Multimodal model list.
AutoModelForSequenceClassificationSequence ClassificationWIPEarly support; interfaces may change.
Diffusers PipelinesDiffusion Generation (T2I, T2V)SupportedSee Diffusion model list.
NeMoAutoModelBiEncoderEmbedding ModelsSupportedSee Embedding model list.
NeMoAutoModelCrossEncoderReranking ModelsSupportedSee Reranking model list.

Release Log

The table below tracks when model support and key features were added across NeMo AutoModel releases. For the full list of tested architectures and example configs, see the LLM, VLM, and Multimodal pages.

ReleaseDateNew ModelsKey Features
0.3.0 (upcoming)—Kimi-VL, Kimi-K25-VL, Gemma 3n, Nemotron-Parse, Qwen3-VL-MoE, Qwen3-Omni, InternVL 3.5, Ministral3, Phi-4-multimodal, Devstral-Small-2, Step-3.5-Flash, Qwen3-Next, Nemotron-3-Nano-30B, FLUX.1-dev, Wan 2.1 T2V, HunyuanVideo 1.5MoE LoRA, expanded VLM coverage, diffusion model training (flow matching)
0.2.0Dec 2025GPT-OSS 20B/120B, Qwen3, Qwen3-MoE, GLM-4/4-MoE, Qwen2.5-VL, Qwen3-VLSingle- and multi-turn tool calling, streaming dataset, QAT for SFT, sequence classification, async DCP checkpointing, MLflow, CP + sequence packing for MoE
0.1.0Oct 2025DeepSeek V3/V3.2, 40+ LLM architectures, Gemma 3 VLMPretraining, knowledge distillation, FP8 (torchao), pipeline parallelism, HSDP, auto pipelining, ColumnMapped dataset
0.1.0a0Sep 2025Initial LLM and VLM support (Llama, Mistral, Qwen2, Gemma, Phi, and more)MegatronFSDP, packed sequences, Triton LoRA kernels

Day-0 Support

  • NeMo AutoModel closely tracks the latest transformers version and updates its dependency regularly.
  • New models released on the Hugging Face Hub may require the latest transformers version, necessitating a package upgrade.
  • We are working on a CI pipeline that automatically bumps the supported transformers version when a new release is detected, enabling even faster day-0 support.

Custom Model Registry

NeMo AutoModel includes a custom model registry that allows teams to:

  • Add custom implementations to extend support to models not yet covered upstream.
  • Provide optimized or faster implementations for specific models while retaining the same AutoModel interface.

Having Issues?

If a model from the Hub doesn’t work as expected, see the Troubleshooting Unsupported Models guide for common issues and solutions.