Seed-OSS-36B-Base
Seed-OSS-36B-Base is an open-weights language model from ByteDance, released August 21, 2025. It has 36B parameters (dense). The context window is 512K tokens.
Seed-OSS is a series of open-source language models developed by ByteDance's Seed Team, designed for powerful long-context, reasoning, agent and general capabilities, and versatile developer-friendly features. We have released Seed-OSS-36B to the open-source community under the Apache-2.0 license.
Check it against your own PC →
Can you run Seed-OSS-36B-Base locally?
The smallest common GPU it runs on is 16 GB of VRAM (with 32 GB of RAM): IQ3_XXS, split GPU / CPU, roughly 6 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 5.2 GB | 13.5 GB | ~2 |
| 8 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 7.2 GB | 11.5 GB | ~2 |
| 12 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 11.2 GB | 7.5 GB | ~4 |
| 16 GB VRAM | Runs, slowly | IQ3_XXS (noticeable loss) | 15.2 GB | 3.5 GB | ~6 |
| 24 GB VRAM | Runs on GPU | Q3_K_M (okay) | 22.6 GB | – | ~31 |
| 32 GB VRAM | Runs on GPU | Q5_K_M (excellent) | 30.6 GB | – | ~40 |
Ready-made quantizations on Hugging Face: NousResearch/Hermes-4.3-36B-GGUF · mradermacher/Seed-OSS-36B-Base-i1-GGUF · mradermacher/Seed-OSS-36B-Base-GGUF · rax4901/Hermes-4.3-36B-GGUF · DevQuasar/ByteDance-Seed.Seed-OSS-36B-Base-GGUF · genevaquant/Hermes-4.3-36B-GGUF
Benchmark results
Epoch AI has not published results for Seed-OSS-36B-Base yet. New models are usually evaluated within days to weeks.
Links
Announcement · Hugging Face: ByteDance-Seed/Seed-OSS-36B-Base
More from ByteDance
Cola2 · Doubao Seed 2.1 Turbo · Doubao Seed 2.1 Pro · Doubao Seed Evolving (Nov 2026) · Doubao Seed 2.0 Lite (Apr 2026) · Doubao-Seed-2.0 Pro · Doubao Seed 2.0 Code · Doubao Seed 2.0 Mini