Can Desktop PC · GeForce RTX 4090 · 64 GB RAM run gpt-oss-120b?
24 GB VRAM + 64 GB RAM · 1008 GB/s memory bandwidth · gpt-oss-120b has 116.8B parameters (5.1B active per token).
Short answer
It loads, but it's borderline even for chat.
Best quantization for chat: MXFP4 (59 GB of weights) on Ollama. Generation: 16–20 tok/s, leaving about 0.0 GB free for your other apps.
How it rates for each workload
Same model and machine, different jobs. Agents and long documents send far bigger prompts than chat, so they need much more speed and memory.
| Workload | Verdict | Quant | Generation | First reply | Context | |
|---|---|---|---|---|---|---|
| Chatwith Open WebUI | Borderline | MXFP4 | 16–20 tok/s | ≈2 s | 8K | Details |
| Coding questionswith Continue | Borderline | MXFP4 | 16–20 tok/s | ≈15 s | 16K | Details |
| Coding in a repositorywith Aider | Borderline | MXFP4 | 14–18 tok/s | ≈45 s | 32K | Details |
| Agentic codingwith OpenCode | Borderline | MXFP4 | 13–18 tok/s | ≈60 s | 32K | Details |
| Long documents (64K)with Open WebUI | Borderline | MXFP4 | 10–13 tok/s | ≈3 min | 64K | Details |
Speeds are estimates calibrated against measured benchmarks on similar hardware. “First reply” is the time to read the workload’s opening prompt. How we calculate
Get it running
Copy-paste commands for the MXFP4 build with a 8K context. Ollama and llama.cpp download the file straight from Hugging Face.
ollama run hf.co/unsloth/gpt-oss-120b-GGUF:F16
Downloads 65.37 GB (F16). Ollama's default context is small: set OLLAMA_CONTEXT_LENGTH=8192 before `ollama serve` for long prompts.
Other models for Desktop PC · GeForce RTX 4090 · 64 GB RAM
The best-rated models for chat on this machine.
- Qwen3.6 35B-A3B240–330 tok/sExcellent
- Gemma 4 26B-A4B160–220 tok/sExcellent
- LFM2.5 24B-A2B360–490 tok/sExcellent
- Qwen3.5 35B-A3B240–330 tok/sExcellent
- GLM-4.7-Flash (30B-A3B)240–330 tok/sExcellent
- gpt-oss-20b≈210 tok/sExcellent
Cheapest machines for gpt-oss-120b
Hardware rated comfortable or better for chat, lowest price first.
- Ryzen AI Max+ 395 mini PC / Framework Desktop · 128 GB$2,000 · 50–70 tok/sExcellent
- Mac Studio M2 Max 96 GB$3,799 · 65–85 tok/sExcellent
- MacBook Pro M3 Max 96 GB$3,999 · 50–65 tok/sExcellent
- Mac Studio M4 Max 128 GB$4,099 · 85–110 tok/sExcellent
- Mac Studio M5 Max 128 GB$4,099 · 80–130 tok/sExcellent
- MacBook Pro M2 Max 96 GB$5,099 · 65–85 tok/sExcellent
Same chip, other memory: 32 GB
Different tool, context size or workload?
These pages use typical settings. The full check models your exact setup.