Can I run Seed OSS 36B Instruct locally?
Seed OSS 36B Instruct by ByteDance-Seed has 36B parameters. At 4-bit with an 8K context it needs about 23.0 GB of GPU memory. It fits on 36 of 65 GPUs and Macs we track.
Memory needed by quant and context length
GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/Seed-OSS-36B-Instruct-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.
| Quant | Quality | 4K ctx | 8K ctx | 32K ctx | 128K ctx |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | rough | 14.4 GB | 15.4 GB | 21.4 GB | 45.4 GB |
| 3-bit (Q3_K_M) • | usable | 18.1 GB | 19.1 GB | 25.1 GB | 49.1 GB |
| 4-bit (Q4_K_M) • | good, most popular | 22.0 GB | 23.0 GB | 29.0 GB | 53.0 GB |
| 5-bit (Q5_K_M) • | very good | 25.5 GB | 26.5 GB | 32.5 GB | 56.5 GB |
| 6-bit (Q6_K) • | near lossless | 29.3 GB | 30.3 GB | 36.3 GB | 60.3 GB |
| 8-bit (Q8_0) • | effectively lossless | 37.5 GB | 38.5 GB | 44.5 GB | 68.5 GB |
| 16-bit (F16/BF16) • | original weights | 69.0 GB | 70.0 GB | 76.0 GB | 100 GB |
Which GPUs and Macs run Seed OSS 36B Instruct?
Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).
Questions
How much VRAM does Seed OSS 36B Instruct need?
About 23.0 GB at 4-bit (Q4_K_M) with an 8K context: 20.3 GB of weights (real GGUF file size), 2.0 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 38.5 GB.
What is the smallest GPU that runs Seed OSS 36B Instruct?
At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 5090 (32 GB).
Can I run Seed OSS 36B Instruct on a Mac?
Yes. At Q4_K_M with 8K context it fits on a Mac M4 Pro 48GB or larger, using the default macOS GPU memory limit.
Model facts
| Hugging Face repo | ByteDance-Seed/Seed-OSS-36B-Instruct |
| Parameters | 36,151,104,512 (36B) |
| Layers / hidden size | 64 / 5120 |
| Attention | GQA; 80 heads, 8 KV heads, head dim 128 |
| KV cache per token (FP16) | 256.0 KiB |
| Max context | 524,288 tokens |
| License | apache-2.0 |
| Released on HF | Aug 20, 2025 |
| Downloads (30d) / likes | 43,432 / 506 |
Sources: Hugging Face model API · config.json (ByteDance-Seed/Seed-OSS-36B-Instruct) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.