Only with CPU offload
gpt-oss 20B at MXFP4 needs 11.6 GB but only 5.3 GB is addressable, so about 58% of the layers would stream from system RAM at roughly 60 GB/s. Expect around 11 tokens per second — usable for batch work, painful for chat.
Ships natively in MXFP4, so the 4-bit weights are the reference weights, not a lossy copy. Fits 16 GB.
Hardware requirements What hardware do I need to run gpt-oss 20B? →At MXFP4 with an 8K context: a 16 GB GPU (RTX 5080) or a 24 GB Mac (M4 Pro). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.
The VRAM budget
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| MXFP4 | 10.8 GB | 11.6 GB | — | ~11 | Reference | 6.3 GB over |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. This model interleaves sliding-window layers (128 tokens, 1 global in 2), which is why its cache barely grows with context.
How to run it
$ llama-server \
-hf openai/gpt-oss-20b:MXFP4 \
-c 8192 -ngl 9
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
Related models on a GeForce GTX 1660 Super
| Model | Total | Tok/s est. | |
|---|---|---|---|
| gpt-oss 120B 117B MoE · same family |
MXFP4 · 61.4 GB | ~5.6 | CPU offload |
| Mistral Small 3.2 24B 23.6B · similar size |
Q4_K_M · 15.1 GB | ~3.5 | CPU offload |
| Devstral Small 2 24B 24B · similar size |
Q4_K_M · 15.3 GB | ~3.4 | CPU offload |
| Gemma 4 26B-A4B 26.5B MoE · similar size |
Q4_K_M · 16.0 GB | ~8.5 | CPU offload |
| Gemma 3 27B 27.4B · similar size |
Q4_K_M · 17.0 GB | ~2.9 | CPU offload |
| Qwen3.8 27B 27.8B · similar size |
Q4_K_M · 16.7 GB | ~3.0 | CPU offload |
| Qwen3.6 27B 27.8B · similar size |
Q4_K_M · 16.7 GB | ~3.0 | CPU offload |
gpt-oss 20B on other hardware
| Device | Total | Tok/s est. | |
|---|---|---|---|
| GeForce RTX 4060 Laptop 7.0 GB usable · 256 GB/s |
11.6 GB | ~13 | CPU offload |
| GeForce RTX 3060 12 GB 10.6 GB usable · 360 GB/s |
11.6 GB | ~31 | CPU offload |
| GeForce RTX 5080 14.4 GB usable · 960 GB/s |
11.6 GB | 120 | Runs great |
| M4 · 16 GB 10.7 GB usable · 120 GB/s |
11.6 GB | — | Won't fit |
| GeForce RTX 4090 22.4 GB usable · 1008 GB/s |
11.6 GB | 126 | Runs great |
| Radeon RX 7900 XTX 22.4 GB usable · 960 GB/s |
11.6 GB | 120 | Runs great |
| M4 Pro · 24 GB 18.0 GB usable · 273 GB/s |
11.6 GB | 38 | Runs great |
| M3 Max · 64 GB 48.0 GB usable · 400 GB/s |
11.6 GB | 56 | Runs great |