Model customization - Amazon SageMaker AI
OverviewChoosing your approachSupported models for fine-tuning
Services or capabilities described in AWS documentation might vary by Region. To see the differences applicable to the AWS European Sovereign Cloud Region, see the AWS European Sovereign Cloud User Guide.

Model customization

Amazon SageMaker AI model customization lets you adapt foundation models to your specific use case by training them on your data. You choose a customization technique (how the model learns), a training type (how much of the model to update), and an infrastructure option (where to run the job).

Overview

Customization techniques define how your model learns from data:

  • SFT (Supervised Fine-Tuning) — Train on labeled prompt-response pairs. See SFT.

  • DPO (Direct Preference Optimization) — Train on preferred vs rejected response pairs. See DPO.

  • RFT (Reinforcement Fine-Tuning) — Optimize via reward signals (RLVR or RLAIF). See RFT.

  • MTRL (Multi-Turn Reinforcement Learning) — Train agents for multi-step agentic tasks. See Multi-turn reinforcement learning.

  • Continuous Customization — Chain techniques sequentially. See Continuous customization.

Training types determine how much of the model to update:

  • LoRA (Low-Rank Adaptation) — Trains a small set of adapter weights. Lower cost, faster training. See LoRA.

  • FFT (Full Fine-Tuning) — Updates all model weights. Higher compute, deeper customization. See FFT.

Infrastructure options control where your training job runs:

  • Serverless — Fully managed, no instance selection. Best for quick experimentation and production jobs without ops overhead. Serverless model customization also supports training session APIs, which give you full control over your training loop without managing GPU infrastructure. See Serverless model customization and Training session APIs.

  • SageMaker AI Training Jobs — Ephemeral jobs where you select instance type. Uses Recipes for pre-configured training. See SageMaker AI Training Jobs.

  • HyperPod — Persistent clusters with fault recovery. Uses Recipes for distributed training on EKS or Slurm. See HyperPod.

Choosing your approach

Before fine-tuning, consider whether a simpler approach meets your needs:

Prompt Engineering

Your model gives acceptable answers but needs better tone, format, or style. No training required — fast iteration. See Prompt engineering for foundation models.

RAG (Retrieval Augmented Generation)

Your model lacks knowledge about your specific domain. Augment with external knowledge sources without retraining. See Retrieval Augmented Generation.

Fine-tuning

Your model needs to consistently learn new behaviors, styles, or domain expertise. Train on your data to change model weights. Continue reading this guide.

Supported models for fine-tuning

The following models are available for customization across serverless, SageMaker AI Training Jobs, and HyperPod infrastructure.

Provider Model Model ID Serverless Training Jobs HyperPod
AlibabaQwen3.8 27Bhuggingface-vlm-qwen3-8-27b✓✓✓
AlibabaQwen3.6 27Bhuggingface-vlm-qwen3-6-27b✓✓✓
AlibabaQwen3.6 A3B 35Bhuggingface-vlm-qwen3-6-35b-a3b✓✓✓
AlibabaQwen3.5 27Bhuggingface-vlm-qwen3-5-27b✓✓✓
AlibabaQwen3.5 9Bhuggingface-vlm-qwen3-5-9b✓✓✓
AlibabaQwen3.5 4Bhuggingface-vlm-qwen3-5-4b✓✓✓
AlibabaQwen3 32Bhuggingface-reasoning-qwen3-32b✓✓✓
AlibabaQwen3 14Bhuggingface-reasoning-qwen3-14b✓✓✓
AlibabaQwen3 8Bhuggingface-reasoning-qwen3-8b✓✓✓
AlibabaQwen3 4Bhuggingface-reasoning-qwen3-4b✓✓✓
AlibabaQwen3 1.7Bhuggingface-reasoning-qwen3-1-7b✓✓✓
AlibabaQwen3 0.6Bhuggingface-reasoning-qwen3-06b✓✓✓
AlibabaQwen2.5 Instruct 72Bhuggingface-llm-qwen2-5-72b-instruct✓✓✓
AlibabaQwen2.5 Instruct 32Bhuggingface-llm-qwen2-5-32b-instruct✓✓✓
AlibabaQwen2.5 Instruct 14Bhuggingface-llm-qwen2-5-14b-instruct✓✓✓
AlibabaQwen2.5 Instruct 7Bhuggingface-llm-qwen2-5-7b-instruct✓✓✓
DeepSeekR1 Distill Qwen 32Bdeepseek-llm-r1-distill-qwen-32b✓✓✓
DeepSeekR1 Distill Qwen 14Bdeepseek-llm-r1-distill-qwen-14b✓✓✓
DeepSeekR1 Distill Qwen 7Bdeepseek-llm-r1-distill-qwen-7b✓✓✓
DeepSeekR1 Distill Qwen 1.5Bdeepseek-llm-r1-distill-qwen-1-5b✓✓✓
DeepSeekR1 Distill Llama 70Bdeepseek-llm-r1-distill-llama-70b✓✓✓
DeepSeekR1 Distill Llama 8Bdeepseek-llm-r1-distill-llama-8b✓✓✓
GoogleGemma 4 31Bhuggingface-vlm-gemma-4-31b-it✓✓✓
GoogleGemma 4 26B A4Bhuggingface-vlm-gemma-4-26b-a4b-it✓✓✓
GoogleGemma 4 E4Bhuggingface-vlm-gemma-4-e4b-it✓✓✓
MetaLlama 3.3 Instruct 70Bmeta-textgeneration-llama-3-3-70b-instruct✓✓✓
MetaLlama 3.2 Instruct 3Bmeta-textgeneration-llama-3-2-3b-instruct✓✓✓
MetaLlama 3.2 Instruct 1Bmeta-textgeneration-llama-3-2-1b-instruct✓✓✓
MetaLlama 3.1 Instruct 8Bmeta-textgeneration-llama-3-1-8b-instruct✓✓✓
Mistral AIMinistral 3 Instruct 2512 BF16 14Bhuggingface-vlm-ministral-3-14b-instruct-2512-bf16✓✓✓
Mistral AIMinistral 3 Instruct 2512 BF16 8Bhuggingface-vlm-ministral-3-8b-instruct-2512-bf16✓✓✓
Mistral AIMinistral 3 Instruct 2512 BF16 3Bhuggingface-vlm-ministral-3-3b-instruct-2512-bf16✓✓✓
NVIDIANemotron 3 Super 120B (A12B)huggingface-llm-nvidia-nemotron-3-super-120b-a12b-bf16✓✓✓
NVIDIANemotron 3 Nano 30B (A3B)huggingface-reasoning-nvidia-nemotron-3-nano-30b-a3b-bf16✓✓✓
NVIDIANemotron 3.5 Lightning 30B (A3B)huggingface-reasoning-nemotron-3-5-lightning-30b-a3b-bf16✓✓✓
OpenAIGPT OSS 120Bopenai-reasoning-gpt-oss-120b✓✓✓
OpenAIGPT OSS 20Bopenai-reasoning-gpt-oss-20b✓✓✓

For Amazon Nova models, see the Nova walkthrough under Serverless model customization.