BestLLMfor EN Your hardware. Your LLM. Your call.
APIOpen data Find my LLM
Model fiche

gpt-oss 120B

By OpenAI · United States

Updated 2026-07-13

chat general reasoning moe
Parameters
117B
License
Apache 2.0
Context
125k
VRAM (Q4)
70 GB
Released
April 2025

Overview

OpenAI's first open-weight return: a 117B MoE with 5.1B active parameters, matching o4-mini quality. Fits a single 80 GB GPU and ships under Apache 2.0.

When to pick this model

  • Production deployments wanting OpenAI quality on owned hardware
  • Reasoning and coding workloads at frontier quality
  • 128k-context document analysis on a single 80 GB GPU
  • Apache-licensed alternative to API-only o4-mini

VRAM requirements by quantization

VRAM REQUIRED (GB)24324880128Q4_K_M70 GBQ5_K_M85 GBQ8_0125 GBFP16234 GB
QuantizationVRAM required
Q4_K_M (recommended)70 GB
Q5_K_M85 GB
Q8_0125 GB
FP16 (no quantization)234 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

In practice, gpt-oss 120B is server-class even at Q4_K_M (70 GB). Stepping up to Q8_0 nearly doubles the footprint to 125 GB, and unquantized FP16 weights take 234 GB — plan your GPU around the Q4 or Q5 figure unless you specifically need the higher fidelity.

Without a GPU, gpt-oss 120B needs roughly 100 GB of system RAM to run on CPU via llama.cpp or Ollama — workable for background jobs, but far slower than GPU inference. Throughput estimates from our compatibility engine: around 12 tokens/sec on entry-level GPUs, on the order of 35 tokens/sec on a mid-range card, and up to 90 tokens/sec on high-end hardware — assuming the chosen quantization fully fits in VRAM.

What hardware do you need

The table below matches gpt-oss 120B to common GPU memory tiers, using the highest-fidelity quantization that fully fits each card class. Spilling layers to system RAM works but costs most of the speed, so size your card to the quantization you actually want to run.

GPU memoryExample cardsBest fit for gpt-oss 120B
8 GBRTX 5070 Laptop, RTX 5060, RTX 5060 Ti 8GBDoes not fit — needs 70 GB at Q4_K_M
12 GBRTX 5070, RTX 5070 Ti Laptop, RTX 4080 LaptopDoes not fit — needs 70 GB at Q4_K_M
16 GBRTX 5080, RTX 4080 Super, Radeon RX 9070 XTDoes not fit — needs 70 GB at Q4_K_M
24 GBRTX 4090, Radeon RX 7900 XTX, RTX 5090 LaptopDoes not fit — needs 70 GB at Q4_K_M
32 GBRTX 5090Does not fit — needs 70 GB at Q4_K_M

Which GPU should you buy to run gpt-oss 120B?

To run gpt-oss 120B locally at Q4, you need ~70 GB of VRAM. The best value for this is a Apple Mac Studio (64+ GB unified memory).

Check Apple Mac Studio price on Amazon →

As an Amazon Associate, BestLLMfor earns from qualifying purchases, at no extra cost to you. It does not influence our independent rankings.

Strengths

  • Matches o4-mini on reasoning and coding benchmarks
  • Apache 2.0 license with full commercial use
  • 128k context out of the box
  • Fits on a single 80 GB accelerator

Limitations

  • Around 70 GB VRAM at Q4 — multi-GPU for higher precision
  • MoE deployment is operationally more complex than dense

Typical workloads

In our catalog grid, gpt-oss 120B is filed under Reasoning, Pro Chat, Agents — the use cases where its size/quality trade-off makes the most sense. Its tags translate to concrete workloads: multi-step reasoning and math-flavoured tasks.

The 125k-token context window covers long chats and mid-sized documents, though very large retrieval workloads will need chunking. The Apache 2.0 license is permissive, so shipping it inside a commercial product raises no special legal questions.

Architecture & training

Architecture: MoE · ~117B total / ~20B active · OpenAI open-source · 128k ctx

Training: OpenAI — first open-weight model released by OpenAI under MIT license.

Verdict

The most consequential open-weight release in years — frontier OpenAI quality on a single GPU under Apache 2.0.

Quick start

ollama run openai/gpt-oss:120b

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.

Similar models worth comparing

Frequently asked questions

How much VRAM does gpt-oss 120B need?

At the recommended Q4_K_M quantization, gpt-oss 120B needs about 70 GB of VRAM. Q8_0 takes 125 GB, and unquantized FP16 weights take 234 GB.

Can gpt-oss 120B run without a GPU?

Yes — with roughly 100 GB of system RAM it runs CPU-only through llama.cpp or Ollama. Expect a fraction of GPU speed, which is fine for background or batch jobs but slow for interactive chat.

What context window does gpt-oss 120B support?

gpt-oss 120B supports a 125k-token context window (128,000 tokens).

Can I use gpt-oss 120B commercially?

Yes. gpt-oss 120B is released under Apache 2.0, a permissive open-source license that allows commercial use, modification and redistribution.

How fast is gpt-oss 120B on consumer hardware?

Our compatibility engine estimates on the order of 35 tokens/sec on a mid-range GPU and up to 90 tokens/sec on high-end cards, assuming the quantization fully fits in VRAM.

Which quantization of gpt-oss 120B should I download first?

Start with Q4_K_M (70 GB) — the standard size/quality sweet spot. Step up to Q5_K_M or Q8_0 only if you have VRAM headroom. It does not fit a single 24 GB consumer card — plan for multi-GPU or server hardware.

Tools

Is gpt-oss 120B the right pick for you?

Compute self-hosted ROI → Back to catalog