| Model profile | ||
|---|---|---|
| Model | Size | Approximate size |
| GPT-OSS 20BModel source: provider131,072 tokens max native context | 20.9Bactive 3.6B | 12.5 GB4-bit quantization |
| Compatible hardware Estimated speeds, not benchmark results: calculated from memory bandwidth and model size. Real results can differ significantly because there is no precise formula for deriving LLM generation speed from hardware specifications alone. | ||||
|---|---|---|---|---|
| Hardware | Memory | Bandwidth | vLLM | Estimated generation |
| DDR4 16GB Computer | 16 GBRAM | 51.2 GB/s | no | 9 tok/s4-bit quantization |
| DDR5 16GB Computer | 16 GBRAM | 89.6 GB/s | no | 16 tok/s4-bit quantization |
| DDR5 32GB Computer | 32 GBRAM | 89.6 GB/s | no | 16 tok/s4-bit quantization |
| Mac mini M4 | 24 GBUnified RAM | 120 GB/s | no | 22 tok/s4-bit quantization |
| RTX 2000 Ada | 16 GBGPU VRAM | 224 GB/s | yes | 40 tok/s4-bit quantization |
| Radeon 8060S 96GB | 96 GBUnified RAM | 256 GB/s | yes | 46 tok/s4-bit quantization |
| M4 Pro | 24 GBUnified RAM | 273 GB/s | no | 49 tok/s4-bit quantization |
| DGX Spark | 128 GBUnified RAM | 273 GB/s | yes | 49 tok/s4-bit quantization |
| RTX 4060 Ti 16GB | 16 GBGPU VRAM | 288 GB/s | yes | 51 tok/s4-bit quantization |
| M5 Pro | 64 GBUnified RAM | 307 GB/s | no | 54 tok/s4-bit quantization |
| RTX 5060 Ti 16GB | 16 GBGPU VRAM | 448 GB/s | yes | 77 tok/s4-bit quantization |
| M5 Max | 128 GBUnified RAM | 614 GB/s | no | 101 tok/s4-bit quantization |
| M3 Ultra | 96 GBUnified RAM | 819 GB/s | no | 128 tok/s4-bit quantization |
| RTX 5080 | 16 GBGPU VRAM | 960 GB/s | yes | 145 tok/s4-bit quantization |
| RTX 3090 Ti | 24 GBGPU VRAM | 1008 GB/s | yes | 151 tok/s4-bit quantization |
| M5 Ultra | 256 GBUnified RAM | 1200 GB/s | no | 172 tok/s4-bit quantization |
| RTX 5090 | 32 GBGPU VRAM | 1792 GB/s | yes | 228 tok/s4-bit quantization |
| RTX PRO 6000 | 96 GBGPU VRAM | 1792 GB/s | yes | 228 tok/s4-bit quantization |
| DGX H200 | 1128 GBGPU HBM3e | 4800 GB/s | yes | 388 tok/s4-bit quantization |
| DGX Station | 748 GBCoherent Memory | 7100 GB/s | yes | 449 tok/s4-bit quantization |
| ET900N G3 | 748 GBCoherent Memory | 7100 GB/s | yes | 449 tok/s4-bit quantization |
| DGX B200 | 1440 GBGPU HBM3e | 8000 GB/s | yes | 466 tok/s4-bit quantization |
| GB200 NVL72 | 13400 GBGPU HBM3e | 8000 GB/s | yes | 466 tok/s4-bit quantization |
| GB300 NVL72 | 20000 GBGPU HBM3e | 8000 GB/s | yes | 466 tok/s4-bit quantization |
Token generation speeds were measured for Gemma 4 E2B, Gemma 4 12B, gpt-oss 20B, Gemma 4 26B, Qwen 3.6 35B, and Qwen 3.8 27B using Ollama on a Geekom A9 Max mini PC with 32 GB RAM.
Meta Muse Glimmer, Qwen3.6 35B-A3B, Qwen3-Coder 30B, and GPT-OSS 20B compete in 10 frozen events on GMKtec EVO-X2 via Ollama on ROCm, judging speed and accuracy across tasks like coding, image generat
Tests Ollama and Llama.cpp running inside Docker on ZimaBoard 2, evaluating speed using Gemma-4 with Vision and GPT-OSS from OpenAI, plus an endurance thermal throttle check.