StepFun model

Step 3.7 Flash Hardware Requirements & Server Compatibility

Step 3.7 Flash is a current StepFun release for frontier multimodal agents, long-context reasoning and visual search. The pinned repository contains approximately 402.731GB of model weights. Our 416GB aggregate across at least 4 GPUs figure is a calculated memory screen, while 768GB aggregate across at least 8 GPUs is the safer starting allowance for deployment testing.

Check all 11 products Open official model card

Model version: stepfun-ai/Step-3.7-Flash
Family and variant: Step 3 · 3.7 Flash
Source version: 5f6244077ac6
Source updated: 3 June 2026

StepFun Step 3.7 Flash stepfun-ai/Step-3.7-Flash

Minimum GPU memory: 416GB aggregate across at least 4 GPUs
Recommended hardware: 768GB aggregate across at least 8 GPUs
Licence: Apache License 2.0
Useful for: Language & reasoning, Coding & agents, Vision & OCR

Memory compatibility is a sizing guide. Test the exact model version and workload before choosing hardware.

Buyer verdict

Where Step 3.7 Flash is a sensible fit

Shortlist Step 3.7 Flash when frontier multimodal agents, long-context reasoning and visual search is the priority and the exact licence and runtime suit the organisation. Choose hardware from the recommended allowance, then measure quality and performance on representative work before purchase.

These figures apply to the named model version. Quantisation, fine-tuning, context length, image resolution, batch size and serving software can materially change the hardware needed.

Hardware requirements

Minimum: 416GB aggregate across at least 4 GPUs
Recommended: 768GB aggregate across at least 8 GPUs

This allowance is derived from the pinned artifact and leaves only limited runtime headroom.

This working allowance creates room for runtime allocations and representative workload testing; it is not a performance benchmark.

Technical specification

Step 3.7 Flash model and hardware facts

Specifications shown for source version 5f6244077ac62e04eec3f320501ff8c2b293373a, updated 3 June 2026.

Architecture: Sparse MoE vision-language model
Parameters: 198B total / approximately 11B active
Context: 262,144 tokens
Modalities: Text + image → text
Repository weights: 402.731GB
Native format: BF16 safetensors

Product compatibility

Step 3.7 Flash compatibility across all 11 GPU systems

Systems that fit without model splitting: 0
Systems needing multi-GPU validation: 4

A system is listed as fitting when its GPU memory meets the requirement shown above. Speed, usable context, batch size and concurrent users still need testing with the final model and software configuration.

First multi-GPU opportunity

Enterprise 768

8 GPUs · 768GB

The first system with enough installed GPU memory when supported software divides the model across its GPUs. We confirm software support, GPU layout, context and performance during sizing.

Explore Enterprise 768

sme workstation

Team 32

Larger-memory system recommended

Per GPU: 32GB
Total GPU memory: 32GB
GPU count: 1

This system provides 32GB per GPU and 32GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

sme workstation

Company 64

Larger-memory system recommended

Per GPU: 32GB
Total GPU memory: 64GB
GPU count: 2

This system provides 32GB per GPU and 64GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

sme workstation

Studio 96

Larger-memory system recommended

Per GPU: 96GB
Total GPU memory: 96GB
GPU count: 1

This system provides 96GB per GPU and 96GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

sme workstation

Studio 192

Larger-memory system recommended

Per GPU: 96GB
Total GPU memory: 192GB
GPU count: 2

This system provides 96GB per GPU and 192GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

pcie rack

Value Rack 128

Larger-memory system recommended

Per GPU: 32GB
Total GPU memory: 128GB
GPU count: 4

This system provides 32GB per GPU and 128GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

pcie rack

Value Rack 256

Larger-memory system recommended

Per GPU: 32GB
Total GPU memory: 256GB
GPU count: 8

This system provides 32GB per GPU and 256GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

pcie rack

Enterprise 384

Larger-memory system recommended

Per GPU: 96GB
Total GPU memory: 384GB
GPU count: 4

This system provides 96GB per GPU and 384GB in total. The model requires 416GB aggregate across at least 4 GPUs.

For this model, explore Enterprise 768 .

pcie rack

Enterprise 768

Multi-GPU route available

Per GPU: 96GB
Total GPU memory: 768GB
GPU count: 8

Total installed memory is sufficient, but the model must be divided across GPUs. Confirm that the serving software supports this GPU layout and test the required context and speed.

pcie rack

H200 1.1TB

Multi-GPU route available

Per GPU: 141GB
Total GPU memory: 1,128GB
GPU count: 8

Total installed memory is sufficient, but the model must be divided across GPUs. Confirm that the serving software supports this GPU layout and test the required context and speed.

frontier partner

Frontier Native 2.3TB

Multi-GPU route available

Per GPU: 288GB
Total GPU memory: 2,304GB
GPU count: 8

Total installed memory is sufficient, but the model must be divided across GPUs. Confirm that the serving software supports this GPU layout and test the required context and speed.

frontier partner

Frontier Rack 20TB

Multi-GPU route available

Per GPU: 288GB
Total GPU memory: 20,736GB
GPU count: 72

Total installed memory is sufficient, but the model must be divided across GPUs. Confirm that the serving software supports this GPU layout and test the required context and speed.

Deployment reality

Strengths, limits and runtime route

What it is good at

Native visual understanding.
Sparse activation limits active compute relative to total size.
256K context.
First-party serving instructions cover four runtime families.

Where to be cautious

The BF16 checkpoint is a multi-GPU deployment.
Model support is recent and exact topology/runtime combinations need validation.
Repository weight size does not include every runtime allocation, KV cache, media encoder, batch or concurrent session.
No GPU Servers benchmark result is claimed until this exact revision has been run under disclosed test conditions.

Serving software

vLLM
SGLang
Transformers
llama.cpp

Before installation

Use source version 5f6244077ac62e04eec3f320501ff8c2b293373a rather than an unversioned latest branch.
Start with vLLM only after checking support for the exact architecture and numerical format.
Use 768GB aggregate across at least 8 GPUs as the procurement starting point; the lower figure is a minimum-memory screen, not a service-level promise.
Record precision, context, batch, concurrency, container digest, driver and measured latency in the acceptance report.

System requirements

GPU layout: The 416GB aggregate across at least 4 GPUs screen assumes supported tensor or pipeline parallel loading. Aggregate memory is not automatically pooled.
Context and cache: 262,144 tokens. Context length, KV cache, batch and concurrent sessions must be tested together; the architecture limit is not a throughput guarantee.
System RAM: Plan system RAM above the 402.731GB repository-weight footprint where model staging or CPU offload is required.
Storage: Reserve at least 1007GB for the pinned weights, runtime cache and one rollback copy; production datasets and logs are additional.
Serving software: First-party material names vLLM, SGLang, Transformers, llama.cpp. Pin the serving version and container digest because support for recent architectures can change quickly.
Representative workload: Acceptance-test frontier multimodal agents, long-context reasoning and visual search at the intended quality, context, batch, concurrency and response-time target.

Performance: speed, usable context and concurrency depend on the selected system, software and workload. No benchmark is quoted on this page.

Commercial and legal boundary

Apache License 2.0

Commercial use: permitted

The official checkpoint is released under Apache 2.0.

Always retain the applicable notices and recheck the live terms for the intended organisation, territory, use and distribution route. Obtain legal advice where required; the official licence governs use.

Read the official licence

Official sources

Technical questions

Step 3.7 Flash deployment FAQ

What GPU memory does Step 3.7 Flash need?

Use 416GB aggregate across at least 4 GPUs as the lower memory screen and 768GB aggregate across at least 8 GPUs as the safer starting allowance. The final requirement changes with runtime, context, batch, concurrency and precision.

Which GPU Servers product should I start with for Step 3.7 Flash?

Use the compatibility table to find systems that meet the recommended allowance. A system that only meets the minimum can load the model in principle but may not meet the required context or response time.

Can Step 3.7 Flash be used commercially?

The official checkpoint is released under Apache 2.0. The linked official licence is authoritative; legal advice may be appropriate for the intended use and distribution route.

Diagram showing an approved request, a local service, an approved store and a policy-controlled data path — A private deployment starts with the permitted data path, access policy and logging boundary. GPU Servers technical illustration.

Diagram showing an approved request, a local service, an approved store and a policy-controlled data path — A private deployment starts with the permitted data path, access policy and logging boundary. GPU Servers technical illustration.

GPU server remote management dashboard with system status, access logs and sensor monitoring panels — Supplier screenshot of the platform management interface. The final management features and access policy depend on the ordered system. OEM supplier reference image.

GPU server remote management dashboard with system status, access logs and sensor monitoring panels — Supplier screenshot of the platform management interface. The final management features and access policy depend on the ordered system. OEM supplier reference image.

Diagram showing approved documents moving through a searchable index to an answer with a source citation — A retrieval workflow should connect each useful answer to approved source material and defined refusal behaviour. GPU Servers technical illustration.

Diagram showing approved documents moving through a searchable index to an answer with a source citation — A retrieval workflow should connect each useful answer to approved source material and defined refusal behaviour. GPU Servers technical illustration.

Diagram of an evidence pack containing an asset schedule, burn-in record, health readings, workload test and admin guide — A complete handover includes the supplied assets, test results, operating instructions and agreed follow-up work. GPU Servers technical illustration.

Diagram of an evidence pack containing an asset schedule, burn-in record, health readings, workload test and admin guide — A complete handover includes the supplied assets, test results, operating instructions and agreed follow-up work. GPU Servers technical illustration.