// StepFun Step · Omni / audio LLM
Step-Audio-AQAA: VRAM & local deployment
136.7B parameters. About 127 GB at 4-bit and 337 GB at BF16 with one 8K-token request.
released 2025-06-11licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | StepFun · stepfun-ai |
|---|---|
| Family / series | StepFun Step · Step-Audio-AQAA |
| Type | Omni / audio LLM |
| Parameters | 136.7B Source: counted from official safetensors weights |
| Context length | — |
| Architecture | 88 layers · GQA |
| Released weights | BF16 · 273 GBdownload |
| Licence | Apache-2.0 · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2025-06-11 (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) | 273 GB | 337 GB | 443 GB | 868 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB |
| FP8 / INT8 / Q8_0 | 145 GB | 196 GB | 302 GB | 727 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 82.8 GB | 127 GB | 233 GB | 659 GB | 4 × RTX PRO 6000 (384GB) | Mac 256GB |
KV cache per request: 8K 35.4 GB · 32K 142 GB · 128K 567 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.
The official card does not name a runtime; see the repository.
04Official repositories & quantised builds
1 official repo(s) from stepfun-ai.
- Step-Audio-AQAA main
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 |
| step3 | 321B / 38B | 247 GB | 2025-07-28 |
06Sources
Data checked 2025-06-12. Parameters, context and licence are copied from the official repository; memory figures are our estimates.