StepFun - Mixture of experts - verified
Run Step 3.7 Flash locally
Step 3.7 Flash is a 201.4B mixture-of-experts (11B active) model from StepFun with a 262K-token context window, released under Apache 2.0 (commercial use is allowed); the smallest listed download is 111.5 GB, the Q4_K_S GGUF build from stepfun-ai/Step-3.7-Flash-GGUF.
Official weights: stepfun-ai/Step-3.7-Flash - modalities: text, image
- Total params
- 201.4B
- Active params
- 11B
- Context
- 262K
- Smallest download
- 111.5 GB
Fits in
Smallest download at most 85% of memoryDownloads
The official weights and every quantization listed for this model, with file sizes from Hugging Face.
| Download | Format | Size | Repo |
|---|---|---|---|
Official weights | BF16 | 402.7 GB | stepfun-ai/Step-3.7-Flash |
Q4_K_SSmallest listed | GGUF | 111.5 GB | stepfun-ai/Step-3.7-Flash-GGUF |
Q8_0 | GGUF | 209.4 GB | stepfun-ai/Step-3.7-Flash-GGUF |
GB is decimal: 1 GB = 10^9 bytes. Add memory for the context and the operating system on top of the file size.
Run it
Copied verbatim from the page linked under each command. Server commands from model cards often assume a multi-GPU node; check the flags before you run them.
SGLang
sglang serve --model-path stepfun-ai/Step-3.7-Flash \
--tp 8 \
--reasoning-parser step3p5 \
--tool-call-parser step3p5 \
--enable-multimodal \
--speculative-algorithm EAGLE \
--speculative-num-steps 3 \
--speculative-eagle-topk 1 \
--speculative-num-draft-tokens 4 \
--enable-multi-layer-eagle \
--trust-remote-code \
--host 0.0.0.0 \
--port 8000vLLM
vllm serve <MODEL_PATH_OR_HF_ID> \
--served-model-name step3p7-flash-bf16 \
--tensor-parallel-size 8 \
--enable-expert-parallel \
--disable-cascade-attn \
--reasoning-parser step3p5 \
--enable-auto-tool-choice \
--tool-call-parser step3p5 \
--speculative_config '{"method": "mtp", "num_speculative_tokens": 3}' \
--trust-remote-codellama.cpp
./llama-cli -m Step3.7_Q4_K_S.gguf -b 2048 -ub 2048 -fa on --temp 1.0 -p "What's your name?"Not sure which runtime to use? Compare local LLM runtimes by platform, API and features.
License
- License
- Apache 2.0
apache-2.0 - Commercial use
- Commercial use: YesCommercial use is allowed.
- Gated download
- No. The files download directly.
Worth knowing
The vendor GGUF needs StepFun's step3.7 branch of llama.cpp and the card lists a minimum of 120 GB of unified memory for local use.
Released (per the announcement).
FAQ
How big is the Step 3.7 Flash download?
The official BF16 weights in stepfun-ai/Step-3.7-Flash are 402.7 GB. The smallest listed download is 111.5 GB, the Q4_K_S GGUF build from stepfun-ai/Step-3.7-Flash-GGUF. Plan for more memory than the file size: the context (KV cache) and the operating system need room too.
Can I use Step 3.7 Flash commercially?
Step 3.7 Flash is released under Apache 2.0, and commercial use is allowed. The license text is at https://www.apache.org/licenses/LICENSE-2.0.
How much memory do I need for Step 3.7 Flash?
By this directory's rule of thumb it fits in 256 GB and up: the smallest listed download, 111.5 GB, is at most 85% of 256 GB, which leaves room for the context and the operating system.
What is the context length of Step 3.7 Flash?
262,144 tokens. A longer context needs more memory for the KV cache, so you can run with a shorter one when memory is tight.
Sources
Last verified from the pages below.
- Hugging Face model card - huggingface.co
- HF API - huggingface.co
- Vendor GGUF repo - huggingface.co
- Step 3.7 Flash announcement - static.stepfun.com
Newsletter
New open weights, every week
One email per week. New model releases, open-weight drops and what they mean for what you build.
Keep exploring
Best local AI models
Our curated picks by hardware tier, with benchmarks and the reasoning.
Model API pricing and specs
First-party prices per 1M tokens when you would rather call an API.
AI benchmark leaderboards
Sourced scores, for when you need to compare quality.
Setup guides
Point your coding tools at the model you picked.