Loops supports a curated set of Hugging Face base models. Each entry below has a verified LoRA configuration on Baseten infrastructure. Coverage expands as new models are validated end to end.Documentation Index
Fetch the complete documentation index at: https://docs.baseten.co/llms.txt
Use this file to discover all available pages before exploring further.
| Model | Inference dtype | Trainer dtype | LoRA dtype | Max sequence length |
|---|---|---|---|---|
Qwen/Qwen3.6-35B-A3B | fp8 | bf16 | bf16 | 131,072 |
Qwen/Qwen3.6-27B | fp8 | bf16 | bf16 | 131,072 |
Qwen/Qwen3.5-9B | bf16 | bf16 | bf16 | 131,072 |
Qwen/Qwen3.5-4B | bf16 | bf16 | bf16 | 131,072 |
Qwen/Qwen3.5-2B | bf16 | bf16 | bf16 | 131,072 |
Qwen/Qwen3.5-0.8B | bf16 | bf16 | bf16 | 131,072 |
Qwen/Qwen3.5-122B-A10B | fp8 | bf16 | bf16 | Contact support |
Qwen/Qwen3.5-397B-A17B | fp8 | bf16 | bf16 | Contact support |
moonshotai/Kimi-K2.6 | nvfp4 | bf16 | bf16 | Contact support |
Qwen/Qwen3-30B-Instruct-2507 | bf16 | bf16 | bf16 | 131,072 |
deepseek-ai/DeepSeek-V4-Pro | nvfp4 | bf16 | bf16 | Contact support |
deepseek-ai/DeepSeek-V4-Flash | fp8 | bf16 | bf16 | Contact support |
zai-org/GLM-5.1 | nvfp4 | bf16 | bf16 | Contact support |
MiniMaxAI/MiniMax-M2.7 | fp8 | bf16 | bf16 | Contact support |
base_model when creating a trainer (POST /v1/trainer_sessions/{session_id}/trainers) or to truss loops push. The trainer dtype controls the precision used for forward, backward, and optimizer steps; the LoRA dtype controls the adapter weights; the inference dtype is what the paired sampling server runs the served checkpoint at.
To request a model not listed here, contact support.