stepfun-ai/Step-5-Preview
archivedmodel
Step 5 Preview is a sparse Mixture-of-Experts (MoE) model with 600B total parameters and 27B active parameters. Mixture-of-experts, 604B total, 27B active (352 experts, 8 per token). Context: 1M tokens (config). License: other.
- Org
- stepfun-ai
- Released
- 2026-09-20
- Status
- archived
- License
- other (restricted)
- Parameters
- 604.3B total, 27.0B active
- Architecture
- MoE · 352 experts, 8/token · Step5ForConditionalGeneration
- Context
- 1,048,576 tokens
- Signal
- 115 HF likes · 1430 downloads · 4 X likes · 3 X post(s)
- Archived
- 2026-10-20 (aged out (older than 30 days); stats as last seen)
Runs on
estimate from params; need = weights + FP16 KV cache for 8K tokens + 1GB. A sizing calculation, not tested.
- Q4 ~400.6 GB (estimated size): no listed device
- Q8 ~680.1 GB (estimated size): no listed device
- FP16 ~1246.7 GB (estimated size): no listed device
Links
First published here 2026-09-20 · updated 2026-10-20 · markdown · JSON