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 with 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
AutoModelForSeq2SeqLMText-to-Text (Encoder-Decoder)SupportedSee the HF compatibility mapping and T5 example recipe
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 ClassificationWork in ProgressEarly support. Interfaces might 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)Not announcedKimi-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 and sequence packing for MoE
0.1.0Oct 2025DeepSeek V3/V3.2, more than 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.
  • The team is developing a CI pipeline that automatically updates the supported transformers version when a new release is detected, enabling 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.

Ready-to-Run Architectures

The following table lists architectures represented by the ready-to-run YAML recipes in this repository. It includes both NeMo-native implementations and models that use the standard Hugging Face implementation path.

This is a practical starting set rather than an exhaustive compatibility list. NeMo AutoModel can also work with additional models supported by the installed version of the Hugging Face transformers library, although models without a checked-in recipe may require some configuration for a particular training setup.

ArchitectureSourceImplementation
BagelForConditionalGenerationNeMo nativenemo_automodel.components.models.bagel.model.BagelForUnifiedMultimodal
BaiChuanForCausalLM (BaichuanForCausalLM)NeMo nativenemo_automodel.components.models.baichuan.model.BaichuanForCausalLM
BailingMoeV2ForCausalLMNeMo nativenemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM
DeepseekV32ForCausalLMNeMo nativenemo_automodel.components.models.deepseek_v32.model.DeepseekV32ForCausalLM
DeepseekV3ForCausalLMNeMo nativenemo_automodel.components.models.deepseek_v3.model.DeepseekV3ForCausalLM
DeepseekV4ForCausalLMNeMo nativenemo_automodel.components.models.deepseek_v4.model.DeepseekV4ForCausalLM
DiffusionGemmaForBlockDiffusionNeMo nativenemo_automodel.components.models.diffusion_gemma.model.DiffusionGemmaForBlockDiffusion
Ernie4_5_MoeForCausalLMNeMo nativenemo_automodel.components.models.ernie4_5.model.Ernie4_5_MoeForCausalLM
Gemma4ForConditionalGenerationNeMo nativenemo_automodel.components.models.gemma4_moe.model.Gemma4ForConditionalGeneration
Glm4MoeForCausalLMNeMo nativenemo_automodel.components.models.glm4_moe.model.Glm4MoeForCausalLM
Glm4MoeLiteForCausalLMNeMo nativenemo_automodel.components.models.glm4_moe_lite.model.Glm4MoeLiteForCausalLM
GlmMoeDsaForCausalLMNeMo nativenemo_automodel.components.models.glm_moe_dsa.model.GlmMoeDsaForCausalLM
GptOssForCausalLMNeMo nativenemo_automodel.components.models.gpt_oss.model.GptOssForCausalLM
HYV3ForCausalLMNeMo nativenemo_automodel.components.models.hy_v3.model.HYV3ForCausalLM
InklingForConditionalGenerationNeMo nativenemo_automodel.components.models.inkling.model.InklingForConditionalGeneration
Kimi-K25-VL (KimiK25ForConditionalGeneration)NeMo nativenemo_automodel.components.models.kimi_k25_vl.model.KimiK25VLForConditionalGeneration
KimiK3ForConditionalGenerationNeMo nativenemo_automodel.components.models.kimi_k3.multimodal.KimiK3ForConditionalGeneration
KimiVLForConditionalGenerationNeMo nativenemo_automodel.components.models.kimivl.model.KimiVLForConditionalGeneration
LagunaForCausalLMNeMo nativenemo_automodel.components.models.laguna.model.LagunaForCausalLM
GritLM (LlamaBidirectionalForSequenceClassification)NeMo nativenemo_automodel.components.models.llama_bidirectional.model.LlamaBidirectionalForSequenceClassification
GritLM (LlamaBidirectionalModel)NeMo nativenemo_automodel.components.models.llama_bidirectional.model.LlamaBidirectionalModel
LlamaForCausalLMNeMo nativenemo_automodel.components.models.llama.model.LlamaForCausalLM
llama-nemotron-embed-1b-v2 (LlamaNemotronVLModel)NeMo nativenemo_automodel.components.models.llama_nemotron_vl.model.LlamaNemotronVLModel
LlavaOneVisionForConditionalGeneration (LLaVAOneVision1_5_ForConditionalGeneration)NeMo nativenemo_automodel.components.models.llava_onevision.model.LLaVAOneVision1_5_ForConditionalGeneration
MiMoV2FlashForCausalLMNeMo nativenemo_automodel.components.models.mimo_v2_flash.model.MiMoV2FlashForCausalLM
MiniMaxM2ForCausalLMNeMo nativenemo_automodel.components.models.minimax_m2.model.MiniMaxM2ForCausalLM
MiniMaxM3SparseForConditionalGenerationNeMo nativenemo_automodel.components.models.minimax_m3_vl.model.MiniMaxM3SparseForConditionalGeneration
Mistral3ForConditionalGenerationNeMo nativenemo_automodel.components.models.mistral4.model.Mistral3ForConditionalGeneration
NemotronHForCausalLMNeMo nativenemo_automodel.components.models.nemotron_v3.model.NemotronHForCausalLM
NemotronParseForConditionalGenerationNeMo nativenemo_automodel.components.models.nemotron_parse.model.NemotronParseForConditionalGeneration
Qwen2_5OmniModelNeMo nativenemo_automodel.components.models.qwen2_5_omni.model.Qwen2_5OmniThinkerForConditionalGeneration
Qwen2ForCausalLMNeMo nativenemo_automodel.components.models.qwen2.model.Qwen2ForCausalLM
Qwen3.5 (Qwen3_5ForConditionalGeneration)NeMo nativenemo_automodel.components.models.qwen3_5.model.Qwen3_5ForConditionalGeneration
Qwen3.5-MoE (Qwen3_5MoeForConditionalGeneration)NeMo nativenemo_automodel.components.models.qwen3_5_moe.model.Qwen3_5MoeForConditionalGeneration
Qwen3ForCausalLMNeMo nativenemo_automodel.components.models.qwen3.model.Qwen3ForCausalLM
Qwen3MoeForCausalLMNeMo nativenemo_automodel.components.models.qwen3_moe.model.Qwen3MoeForCausalLM
Qwen3VLForConditionalGenerationNeMo nativenemo_automodel.components.models.qwen3_vl.model.Qwen3VLForConditionalGeneration
Qwen3VLForConditionalGeneration (Qwen3VLMoeForConditionalGeneration)NeMo nativenemo_automodel.components.models.qwen3_vl_moe.model.Qwen3VLMoeForConditionalGeneration
Step3p5ForCausalLMNeMo nativenemo_automodel.components.models.step3p5.model.Step3p5ForCausalLM
Step3p7ForConditionalGenerationNeMo nativenemo_automodel.components.models.step3p7.model.Step3p7ForConditionalGeneration
CohereForCausalLMHugging Facetransformers
FalconH1ForCausalLMHugging Facetransformers
Gemma2ForCausalLMHugging Facetransformers
Gemma3ForConditionalGenerationHugging Facetransformers
Gemma 3n (Gemma3nForConditionalGeneration)Hugging Facetransformers
Gemma4UnifiedForConditionalGenerationHugging Facetransformers
GlmForCausalLMHugging Facetransformers
GraniteForCausalLMHugging Facetransformers
InternVLChatModelHugging Facetransformers
MistralForCausalLMHugging Facetransformers
MixtralForCausalLMHugging Facetransformers
NemotronFlashForCausalLMHugging Facetransformers
OLMo2ForCausalLM (Olmo2ForCausalLM)Hugging Facetransformers
Phi3ForCausalLMHugging Facetransformers
Phi4MMForCausalLMHugging Facetransformers
Qwen2_5VLForConditionalGeneration (Qwen2_5_VLForConditionalGeneration)Hugging Facetransformers
Starcoder2ForCausalLMHugging Facetransformers
T5ForConditionalGenerationHugging Facetransformers

Having Issues?

If a model from the Hub does not work as expected, see Troubleshooting for common issues and solutions.