// StepFun Step · Vision-language
step3: VRAM & local deployment
321B total · 38B active (MoE) parameters. About 247 GB at 4-bit and 740 GB at BF16 with one 8K-token request.
released 2025-07-28licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | StepFun · stepfun-ai |
|---|---|
| Family / series | StepFun Step · step3 |
| Type | Vision-language (image-text-to-text) |
| Parameters | 321B total · 38B active (MoE) Source: official model card“321B total parameters and 38B active” |
| Context length | — |
| Architecture | 61 layers · GQA · MoE |
| Released weights | BF16 · 642 GBdownload |
| Licence | Apache-2.0 · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2025-07-28 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Total @ 8K | Total @ 32K | Total @ 128K | Smallest GPU | Mac |
|---|---|---|---|---|---|---|
| BF16 / FP16 (original) | 642 GB | 739 GB | 838 GB | 1.23 TB | 8 × H200 (1,128GB) | — |
| FP8 / INT8 / Q8_0 | 341 GB | 408 GB | 507 GB | 900 GB | 8 × H100 (640GB) | — |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 195 GB | 247 GB | 346 GB | 739 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB |
KV cache per request: 8K 32.8 GB · 32K 131 GB · 128K 524 GB
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- vLLM — mentioned in the official model card
- SGLang — mentioned in the official model card
- Transformers — mentioned in the official model card
04Official repositories & quantised builds
2 official repo(s) from stepfun-ai.
05Other StepFun Step models
All 12 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| Step-5-Preview | 600B / 27B | 412 GB | 2026-09-20 |
| Step-3.7-Flash | 198B / 11B | 136 GB | 2026-05-23 |
| Step-3.5-Flash | 196B / 11B | 135 GB | 2026-02-01 |
| Step-Audio-R1.1 | 33.5B | 25.0 GB | 2026-01-14 |
| Step3-VL-10B | 10.2B | 8.5 GB | 2026-01-13 |
| Step-Audio-EditX | 3.53B | 6.1 GB | 2025-10-29 |
| Step-Audio-2-mini | 8.31B | 6.5 GB | 2025-08-28 |
| Step-Audio-AQAA | 136.7B | 127 GB | 2025-06-11 |
06Sources
Data checked 2026-01-29. Parameters, context and licence are copied from the official repository; memory figures are our estimates.