Supported base models

Hugging Face base models Loops accepts, with sequence-length limits.

Each row below is a Hugging Face repo ID you can pass as base_model when starting a Loops run, along with the maximum supported sequence length. Baseten adds rows as it validates new models end to end.

Models

Model	Max sequence length
Qwen/Qwen3-0.6B	8,192
Qwen/Qwen3-4B-Instruct-2507	40,960
Qwen/Qwen3-8B	131,072
Qwen/Qwen3-30B-A3B-Instruct-2507	131,072
Qwen/Qwen3.5-0.8B	131,072
Qwen/Qwen3.5-2B	131,072
Qwen/Qwen3.5-4B	131,072
Qwen/Qwen3.5-9B	131,072
Qwen/Qwen3.5-27B	131,072
Qwen/Qwen3.5-35B-A3B	131,072
Qwen/Qwen3.5-122B-A10B	131,072
Qwen/Qwen3.5-397B-A17B	131,072
Qwen/Qwen3.6-27B	131,072
Qwen/Qwen3.6-35B-A3B	131,072
deepseek-ai/DeepSeek-V4-Flash	131,072
moonshotai/Kimi-K2.6	131,072
moonshotai/Kimi-K2.7-Code	131,072
nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16	262,144
zai-org/GLM-5.2-FP8	65,536

List supported models

Query the /v1/loops/capabilities endpoint for the current list of base models and their maximum sequence lengths:

Request

curl https://api.baseten.co/v1/loops/capabilities \
  -H "Authorization: Bearer $BASETEN_API_KEY"

The endpoint lists the base models your workspace has access to, so its response is the source of truth for what you can pass as base_model, even where it differs from the table above. See GET /v1/loops/capabilities for the full route reference.

Pass a model to Loops

Pass the table value verbatim as base_model through any of the following entry points:

• The Python SDK, using tinker.ServiceClient.create_lora_training_client(base_model=...). See the Loops quickstart.
• The HTTP API, using POST /v1/loops/runs.
• The CLI, using truss loops push <base_model>, which provisions a session, run, and paired sampler in one call.

The minimal HTTP call provisions a run and its paired sampler against an existing session. Replace 2qjl22w with the session.id returned by POST /v1/loops/sessions:

curl --request POST \
  --url https://api.baseten.co/v1/loops/runs \
  --header "Authorization: Bearer $BASETEN_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "session_id": "2qjl22w",
    "base_model": "Qwen/Qwen3.5-9B"
  }'

For the full request body, response shape, and an interactive playground, see POST /v1/loops/runs in the Loops API reference.

Request a model

To request a base model that isn’t listed, contact support.