Skip to content
English

Switch theme

// ByteDance Seed · Text LLM

Seed-OSS-36B-Instruct: VRAM & local deployment

36.2B parameters. About 26.8 GB at 4-bit and 82.2 GB at BF16 with one 8K-token request.

released 2025-08-20licence: Commercial OKmemory figures are estimates

01Specs

PublisherByteDance Seed · ByteDance-Seed
Family / seriesByteDance Seed · Seed-OSS
TypeText LLM (text-generation)
Parameters36.2B Source: counted from official safetensors weights
Context length512K tokens (524,288)max_position_embeddings in config.json
Architecture64 layers · GQA
Released weightsBF16 · 72.3 GBdownload
LicenceApache-2.0 · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed.
Released2025-08-20 (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.

PrecisionWeightsTotal @ 8KTotal @ 32KTotal @ 128KSmallest GPUMac
BF16 / FP16 (original)72.3 GB 82.2 GB88.6 GB114 GB RTX PRO 6000 96GBMac 128GB
FP8 / INT8 / Q8_038.4 GB 44.9 GB51.3 GB77.1 GB 2 × RTX 5090 (64GB)Mac 128GB
4-bit (Q4_K_M / AWQ / GPTQ-Int4)21.9 GB 26.8 GB33.2 GB59.0 GB RTX 5090 32GBMac 64GB

KV cache per request: 8K 2.1 GB · 32K 8.6 GB · 128K 34.4 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.

Which inference engine to choose →

04Official repositories & quantised builds

2 official repo(s) from ByteDance-Seed.

05Other ByteDance Seed models

All 9 →
ModelParams4-bit totalReleased
Seed-X-Instruct-7B7.51B6.6 GB2025-07-16
Seed-X-PPO-7B7.51B6.6 GB2025-07-16
SeedVR-3B3B2.5 GB2025-06-11
SeedVR2-7B7B5.2 GB2025-06-11
SeedVR2-3B3B2.5 GB2025-06-10
BAGEL-7B-MoT14.7B10.8 GB2025-05-19
Seed-Coder-8B-Instruct8.25B7.1 GB2025-04-27
UI-TARS-1.5-7B8.29B6.5 GB2025-04-16

06Sources

Data checked 2025-08-26. Parameters, context and licence are copied from the official repository; memory figures are our estimates.

Data updated: Self-host open models(UTC+8)