Skip to content
View original post on X: StepFunOfficial
Published
34/100

StepFun's Step-5-preview gets Day-0 support across inference and cloud platforms

AIStepFun says its Step-5-preview model can be served through an OpenAI-compatible API with vLLM or SGLang, or run offline with Transformers, and it is adapted for multiple chip platforms and inference frameworks. NVIDIA is offering Day-0 support, including a build.nvidia.com preview, NeMo AutoModel fine-tuning, a model-free NIM deployment, and Blackwell support.
StepFunVerified on XOfficial
@StepFun_ai
View on X

Part of a thread · earlier post

Run it your way:

  • Serve an OpenAI-compatible API with vLLM or SGLang
  • Run offline inference with Transformers
  • Adapted for multiple chip platforms and inference frameworks

Day-0 support from @NVIDIAAI:
try it on https://build.nvidia.com/stepfun-ai/step-5-preview,
fine-tune with NeMo AutoModel https://github.com/NVIDIA-NeMo/Automodel/blob/main/docs/model-coverage/llm/stepfun-ai/step-5-preview.mdx,
deploy with model-free NIM https://catalog.ngc.nvidia.com/orgs/nim/nvidia/containers/vllm-model-free-nim-experimental,
and run on Blackwell.

Also available on Day 0 with @FireworksAI_HQ, @togethercompute, @modal, @DeepInfra and @FeatherlessAI.

Source details

StepFun · x.comPublished

Original title: Run it your way:

Topics

NVIDIAOpen-source ecosystemDeployment & engineeringChina AIModel releasesProduct updates