Skip to content

Best GPU for gpt-oss-120b: VRAM Requirements

gpt-oss-120b needs about 257.6 GB of GPU memory in BF16 and 64.9 GB at 4 bits. See which GPUs fit it for inference and fine-tuning.

gpt-oss-120b needs about 257.6 GB of VRAM in FP16, 129.1 GB in FP8 and 64.9 GB at 4 bits for inference with an 8,192-token context. In FP16 it fits on one NVIDIA B300 288GB; a 4-bit build fits on one NVIDIA A100 PCIe 80GB. Its weights are published in 4-bit MXFP4, so the 4-bit figure is the closest to the checkpoint.

gpt-oss-120b is a mixture-of-experts language model with 116.8B total and 5.1B active parameters per token from OpenAI, released in August 2025, with a 131,072-token context window and published under the Apache 2.0. Its weights alone take 233.6 GB in BF16 and 58.4 GB at 4 bits per parameter; the tables below add the KV cache and runtime memory, then list the GPUs that hold each setup.

Recommended GPUs for gpt-oss-120b

gpt-oss-120b VRAM requirements by precision

One request with an 8,192-token context, KV cache in BF16, plus a 10 % margin for activations and runtime buffers.

PrecisionWeightsKV cacheTotal for inference
BF16 / FP16233.6 GB0.6 GB257.6 GB
FP8116.8 GB0.6 GB129.1 GB
INT4 (closest to published MXFP4)58.4 GB0.6 GB64.9 GB

gpt-oss-120b VRAM by task

Fine-tuning rows assume a short sequence, batch 1 and gradient checkpointing; longer sequences add activation memory.

TaskGPU memory
Inference at the full 131,072-token context (FP8)139.1 GB
QLoRA (4-bit base)75.9 GB
LoRA (BF16 base)303.7 GB
Full fine-tuning (BF16, AdamW)2,055.7 GB

GPUs that fit gpt-oss-120b

Number of cards whose combined memory holds each setup, assuming the model is split across them with tensor or pipeline parallelism. A dash means more than eight cards.

GPUMemoryBF16 inference (257.6 GB)FP8 inference (129.1 GB)INT4 inference (64.9 GB)QLoRA (75.9 GB)LoRA (303.7 GB)
NVIDIA GeForce RTX 5080 16GB16 GB——8 GPUs8 GPUs—
NVIDIA GeForce RTX 4090 24GB24 GB—8 GPUs4 GPUs4 GPUs—
NVIDIA L4 24GB24 GB—8 GPUs4 GPUs4 GPUs—
NVIDIA GeForce RTX 5090 32GB32 GB—8 GPUs4 GPUs4 GPUs—
NVIDIA Tesla V100 32GB32 GB—8 GPUs4 GPUs4 GPUs—
NVIDIA A100 PCIe 40GB40 GB8 GPUs4 GPUs2 GPUs2 GPUs8 GPUs
NVIDIA L40S 48GB48 GB8 GPUs4 GPUs2 GPUs2 GPUs8 GPUs
NVIDIA Quadro RTX 8000 48GB48 GB8 GPUs4 GPUs2 GPUs2 GPUs8 GPUs
NVIDIA RTX 6000 Ada 48GB48 GB8 GPUs4 GPUs2 GPUs2 GPUs8 GPUs
NVIDIA RTX A6000 48GB48 GB8 GPUs4 GPUs2 GPUs2 GPUs8 GPUs
NVIDIA A100 PCIe 80GB80 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA A100 SXM 80GB80 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA H100 PCIe 80GB80 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA H100 SXM 80GB80 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA RTX PRO 6000 Blackwell 96GB96 GB4 GPUs2 GPUs1 GPU1 GPU4 GPUs
NVIDIA H200 141GB141 GB2 GPUs1 GPU1 GPU1 GPU4 GPUs
NVIDIA B200 180GB180 GB2 GPUs1 GPU1 GPU1 GPU2 GPUs
NVIDIA B300 288GB288 GB1 GPU1 GPU1 GPU1 GPU2 GPUs
NVIDIA GB300 NVL288 GB1 GPU1 GPU1 GPU1 GPU2 GPUs

What is gpt-oss-120b?

gpt-oss-120b is OpenAI's larger open-weight reasoning model, a mixture-of-experts model with 5.1 billion active parameters per token. With its MXFP4 expert weights, OpenAI states that it runs on a single 80 GB GPU such as the H100.

Developer
OpenAI
Released
August 2025
Parameters
116.8B total, 5.1B active per token
Context window
131,072 tokens
License
Apache 2.0

Expert weights are MXFP4 (about 4.25 bits per parameter) and attention weights BF16, so the 4-bit row is the closest to the published checkpoint, which takes about 65 GB, more than the 4-bit weight estimate. Half of the layers use a 128-token sliding window, which keeps the real KV cache below this estimate.

Source: gpt-oss-120b model card (as of November 17, 2026).

Frequently asked questions

How much VRAM does gpt-oss-120b need?
For inference with an 8,192-token context and one request, gpt-oss-120b needs about 257.6 GB in BF16, 129.1 GB in FP8, 64.9 GB in INT4, including a 10 % runtime margin. Real usage depends on the inference engine and its settings.
Can gpt-oss-120b run on a single GPU?
Yes. In BF16 it needs about 257.6 GB, which fits one NVIDIA B300 288GB (288 GB).
How much GPU memory does fine-tuning gpt-oss-120b take?
QLoRA (4-bit base): about 75.9 GB; LoRA (BF16 base): about 303.7 GB; Full fine-tuning (BF16, AdamW): about 2,055.7 GB. Longer training sequences and larger batches add activation memory on top.
What is the best GPU for gpt-oss-120b?
Best for inference: 1 GPU × NVIDIA A100 PCIe 80GB (4-bit weights, 64.9 GB at 8,192 tokens). Best for fine-tuning: 2 GPUs × NVIDIA B200 180GB (LoRA on BF16 weights, 303.7 GB). Budget pick: 1 GPU × NVIDIA RTX PRO 6000 Blackwell 96GB (Smallest card class that holds a 4-bit build, 64.9 GB). Check live offers before booking, as availability changes.
Can gpt-oss-120b run on an RTX 4090?
Not on one card: even a 4-bit build needs about 64.9 GB, so it takes 4 RTX 4090 cards (96 GB combined) with the model split across them.
Can gpt-oss-120b run on an RTX 5090?
Not on one card: even a 4-bit build needs about 64.9 GB, so it takes 4 RTX 5090 cards (128 GB combined) with the model split across them.

Estimates use the formulas of the LLM VRAM calculator (decimal GB). Real usage depends on the inference engine, quantization format and memory fragmentation: test on the target GPU before a long rental.

Explore the live market

Every offer on the live market comes from a reviewed supplier and is sold by EverXCloud Inc.