Can You Run gpt-oss-120B? VRAM Requirements
gpt-oss-120B has 117B total parameters (MoE, 5.1B active per token). Running it at MXFP4 with 16K context takes roughly 63.5 GB of memory (57.8 GB weights + 1.1 GB KV cache + runtime overhead). The minimum viable hardware is about A100 / H100 80GB. All numbers are planning estimates, verified 2026-07-27.
VRAM by quantization (16K context)
| Quantization | Weights | Total needed | Minimum hardware |
|---|---|---|---|
| FP16 / BF16 | 218 GB | 236 GB | Mac M3 Ultra 512GB |
| FP8 | 109 GB | 119 GB | H200 141GB |
| Q8_0 (GGUF) | 116 GB | 126 GB | H200 141GB |
| Q6_K (GGUF) | 89.4 GB | 97.6 GB | Mac 128GB unified |
| Q5_K_M (GGUF) | 77.4 GB | 84.7 GB | Mac 128GB unified |
| MXFP4recommended | 57.8 GB | 63.5 GB | A100 / H100 80GB |
| Q4_K_M (GGUF) | 66.5 GB | 72.9 GB | A100 / H100 80GB |
| Q3_K_M (GGUF) | 51.2 GB | 56.4 GB | Mac 64GB unified |
| Q2_K (GGUF) | 38.1 GB | 42.3 GB | Mac 64GB unified |
How context length changes memory (MXFP4)
| Context | KV cache | Total needed |
|---|---|---|
| 4K | 0.3 GB | 62.7 GB |
| 16K | 1.1 GB | 63.5 GB |
| 64K | 4.5 GB | 66.9 GB |
| 128K | 9.0 GB | 71.4 GB |
KV cache grows linearly with context; FP8 KV cache halves it again (try it in the calculator below).
Customize the estimate
LLM VRAM Calculator — Can I Run It?
Estimate how much GPU memory (VRAM) you need to run open-weight LLMs like Kimi K3, DeepSeek R1, Qwen3, or Llama locally. Pick a model, quantization, and context length — the calculator adds up weights, KV cache, and runtime overhead, then shows which GPUs or Macs can fit it. Runs entirely in your browser: no upload, no signup.
117B total params · 5.1B active (MoE) · max 128K context
Native format of Kimi K3 / gpt-oss · ~0.53 bytes/param
Estimated memory needed
71.4 GB
Will it fit?
Assumes ~92% of device memory is usable. Multi-GPU counts are for tensor/pipeline parallel serving (vLLM, SGLang); Apple Silicon uses unified memory via llama.cpp or MLX.
FAQ
How much VRAM does gpt-oss-120B need?
About 63.5 GB at the recommended MXFP4 quantization with 16K context: 57.8 GB for weights, 1.1 GB for KV cache, plus runtime overhead. Longer context grows the KV cache.
What is the minimum hardware for gpt-oss-120B?
Roughly A100 / H100 80GB, assuming 92% of device memory is usable. Lower-end setups can try more aggressive quantization (Q3/Q2) at a noticeable quality cost.
Can I run gpt-oss-120B on a Mac?
Yes. At MXFP4 with 16K context it needs about 63.5 GB, which fits the unified memory of a Mac M3 Ultra 512GB via llama.cpp or MLX.
Which quantization should I use for gpt-oss-120B?
MXFP4 is the model's native/recommended format — Native format of Kimi K3 / gpt-oss. Pick Q4_K_M for a smaller footprint or FP8 / Q8_0 for near-lossless quality.