{
 "slug": "Seed-OSS",
 "title": "Seed-OSS",
 "type": "model",
 "short_desc": "ByteDance Seed's August 2025 open-weight 36B dense language model family, notable for a long native context window and a user-controllable thinking budget.",
 "categories": [
  "ByteDance models",
  "Chinese AI labs",
  "Open-weight models",
  "2025 model releases",
  "Models"
 ],
 "infobox": {
  "Developer": "ByteDance (Seed team)",
  "Family": "Seed (ByteDance)",
  "Announced": "August 2025",
  "Architecture": "Dense decoder-only Transformer",
  "Parameters": "36 billion (developer-reported)",
  "Context window": "512,000 tokens (developer-reported, native)",
  "Post-training": "Instruction tuning and reinforcement learning (Instruct variant)",
  "License / access": "Apache 2.0; open weights on Hugging Face and GitHub",
  "Paper/report": "ByteDance Seed, Seed-OSS release notes and model cards (August 2025)"
 },
 "infobox_links": {
  "Developer": [
   "ByteDance"
  ],
  "Family": [
   "ByteDance"
  ],
  "Architecture": [
   "Transformer_(architecture)"
  ],
  "Post-training": [
   "Instruction_tuning"
  ],
  "License / access": [
   "Hugging_Face"
  ],
  "Paper/report": [
   "ByteDance"
  ]
 },
 "aliases": [],
 "family": "",
 "org": "ByteDance",
 "registry_status": "stub",
 "words": 335,
 "references": 3,
 "outbound": [
  "ByteDance",
  "Chinese_AI_labs",
  "DeepSeek-R1",
  "Doubao",
  "GLM-4.5",
  "Hugging_Face",
  "Instruction_tuning",
  "Kimi_K2",
  "Large_language_model",
  "Mixture_of_experts",
  "OpenAI",
  "Qwen3",
  "Seedream",
  "Test-time_compute",
  "Transformer_(architecture)",
  "gpt-oss"
 ],
 "inbound": [
  "ByteDance",
  "Chinese_AI_labs",
  "Doubao",
  "Mixture_of_experts",
  "Seedream",
  "Stated_missions_of_AI_labs"
 ],
 "url": "wiki/Seed-OSS.html",
 "built": "2026-07-24"
}