The smallest-memory cataloged machine that runs TinyLlama on the GPU is a AMD Radeon RX 7600 · 8GB VRAM. Here are the rigs that fit it, smallest memory first.
| machine | best fit | verdict | speed |
|---|---|---|---|
| AMD Radeon RX 7600 · 8GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 55–90 tok/s |
| NVIDIA GeForce RTX 3070 · 8GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 90–150 tok/s |
| NVIDIA GeForce RTX 4060 · 8GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 55–90 tok/s |
| Apple M1 · 8GB | 1.1B FP16 | ✅ Runs (tight) | 18–30 tok/s |
| Apple M2 · 8GB | 1.1B FP16 | ✅ Runs (tight) | 25–45 tok/s |
| Apple M3 · 8GB | 1.1B FP16 | ✅ Runs (tight) | 25–45 tok/s |
| NVIDIA GeForce RTX 3080 · 10GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 150–250 tok/s |
| AMD Radeon RX 6700 XT · 12GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 70–120 tok/s |
| AMD Radeon RX 7700 XT · 12GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 80–130 tok/s |
| NVIDIA GeForce RTX 3060 · 12GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 70–120 tok/s |
| NVIDIA GeForce RTX 4070 · 12GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 100–170 tok/s |
| AMD Radeon RX 6800 XT · 16GB VRAM | 1.1B FP16 | ✅ Runs comfortably | 95–160 tok/s |
Already have a machine? See TinyLlama across popular rigs → or check your exact rig →
Planning a whole workflow? Rigs for a coding agent · local chat · reasoning · local RAG setup.
Verdicts use the same GQA-aware fit math as the CLI — a AMD Radeon RX 7600 verdict here matches what runlocal tells you on the machine.