community pool / model answer
Llama-3.1-8B-Instruct
chat · meta-llama/Llama-3.1-8B-Instruct
76 tok/s · model median · 8 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| GTX 1080 Ti 11GB | 4-bit | 52.8 tok/s | reported | - | 7.7 GB | 8,192 | llama.cpp |
| GTX 1080 Ti 11GB | 5-bit | 38.1 tok/s | reported | - | 8.4 GB | 8,192 | llama.cpp |
| RTX 2080 8GB | 4-bit | 76 tok/s | reported | - | 4.9 GB | 2,048 | ollama |
| RTX 3060 12GB | 4-bit | 76 tok/s | reported | - | 4.9 GB | 2,048 | ollama |
| RTX 3080 12GB | 4-bit | 134 tok/s | reported | - | 5.1 GB | 2,048 | ollama |
| RTX3060+RTX2080-pooled-20GB 20GB | 4-bit | 105 tok/s | reported | - | 5 GB | 2,048 | ollama |
Context tested
The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.