LocalClaw/LLM/Step 3.5 Flash
Open-weight MoE

Step 3.5 Flash

Catalogue summary: StepFun open-weight sparse MoE with 196.81B total parameters and about 11B active per token. The verified GGUF currently provides Q4_K at roughly 118.7 GB, requiring a high-memory workstation. Apache 2.0.

Repository editorial metadata; verify comparative claims in the linked upstream material.

192 GB catalogue minimum192 GB RAMQ4_KCoding assistant
Choose an app

Only verified options can be opened. Unsupported apps are clearly marked.

Compare models
LM StudioNot available for this model
Open in UnslothLaunches Unsloth Desktop on this model's download page
OllamaNot available for this model
Open on Hugging FaceFiles, licence and available downloads
llama.cppNot available for this model
Use with LocalClawOptional workspace after the model is installed

Desktop app links require the app to be installed. If nothing opens, LocalClaw will show app-download and model-file fallbacks.

Parameters
196.81B (11B active, MoE)
Minimum RAM
192 GB
Model size
118.7 GB
Quantization
Q4_K

Can Step 3.5 Flash run locally?

Step 3.5 Flash has a catalogue minimum of 192 GB RAM with Q4_K. Actual memory use and speed vary by context length, runtime, backend and system headroom.

Treat step-3.5-flash as a server-grade catalogue target. Use the verified artefact with a compatible multi-GPU or distributed runtime and follow its upstream instructions; this is not a one-click desktop LM Studio recommendation.

Model sourcePublic GGUF repository (verified 2026-09-07): ggml-org/Step-3.5-Flash-GGUF
chatcodepowerreasoningmoe

Deployment path

01
Check RAM fitServer-grade target. Plan for 192 GB class multi-GPU memory.
02
Load the modelUse step-3.5-flash only with a compatible server-grade or distributed runtime; confirm the exact artefact and upstream instructions first.
03
Control locallyUse LocalClaw to manage models, agents, chat, channels and scheduled OpenClaw work.

Catalogue record

  • Family: step
  • Parameters: 196.81B (11B active, MoE)
  • Recommended quantization: Q4_K
  • Catalogue minimum RAM: 192 GB
  • Catalogue model size: 118.7 GB
  • Tags: chat, code, power, reasoning, moe

Practical limits

  • Catalogue RAM is a minimum estimate, not a guarantee for every context length or runtime.
  • Speed and memory use vary by quantization, backend, context length and system headroom.
  • Verify architecture, licence and usage restrictions in the linked upstream material before deployment.

Catalogue tags

  • chat
  • code
  • power
  • reasoning
  • moe

Capability profile

Repository catalogue ratings used by LocalClaw's editorial rubric. They are not a standardized third-party benchmark.

speed
8
quality
7
coding
7
reasoning
7

Technical notes

Developer
See upstream repository
License
See upstream repository
Context window
See upstream repository
Architecture
See upstream repository

This model fits these next steps

Hardware fit is based on LocalClaw's RAM tier, model size and quantization metadata. Always leave memory headroom for your OS and runtime.

Very large memoryMac Studio Ultra classCheck model size firstNVIDIA GB10 / server optionsMore practical alternativesCompare smaller models

Related catalogue entries

Linked mechanically by family, shared tags and nearby RAM tier; this is not a quality ranking.

Qwen 3 MoE (235B/22B active) 235B (22B active)Ling Flash Base 2.0 106B (6.1B active, MoE)MiniMax M2 (230B MoE) 230B (10B active)MiniMax M2.1 230B (10B active, MoE)DeepSeek V4 Flash (284B MoE) 284B (13B active)GLM-5.3-Flash 320B (18B active, MoE)LongCat-Flash-Lite 68.5B (3-4.5B active, MoE)Qwen3-Coder-Next 80B (3B active, MoE)

Where to go next