6 builds of Gemma 3 1B measured on real machines — different packagers, precisions and engines, each a different file with its own numbers. The best each engine proved is first; every build has its own page with the machines and signatures behind it.
trained 32K · proved 8K context proved by MaziyarPanahi Q3_K_S
13.6 tok/s peak, by MaziyarPanahi Q3_K_S
proved by at least one build: Structured output · Sustained output · Writes code · Follows instructions
| build | publisher | precision | engine | whose work | context proved | best tok/s | machines |
|---|---|---|---|---|---|---|---|
| hf.co/MaziyarPanahi/gemma-3-1b-it-GGUF:Q3_K_S | MaziyarPanahi | Q3_K_S | ollama | quantizer | 8,192 | 13.6 | 1 |
| hf.co/lmstudio-community/gemma-3-1b-it-GGUF:Q3_K_L | lmstudio-community | Q3_K_L | ollama | quantizer | 8,192 | 13.1 | 1 |
| hf.co/MaziyarPanahi/gemma-3-1b-it-GGUF:Q2_K | MaziyarPanahi | Q2_K | ollama | quantizer | 8,192 | 13.0 | 1 |
| hf.co/ggml-org/gemma-3-1b-it-GGUF:Q4_K_M | ggml-org | Q4_K_M | ollama | quantizer | 8,192 | 12.6 | 1 |
| hf.co/lmstudio-community/gemma-3-1b-it-GGUF:Q6_K | lmstudio-community | Q6_K | ollama | quantizer | 8,192 | 10.8 | 1 |
| hf.co/ggml-org/gemma-3-1b-it-GGUF:Q8_0 | ggml-org | Q8_0 | ollama | quantizer | 8,192 | 9.7 | 1 |