2 builds of Gemma 4 31B 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 256K · proved 128K context proved by unsloth Q4_K_M
3.4 tok/s peak, by unsloth Q4_K_M
proved by at least one build: Tool calling · Tool loop finished · Faithful to tool results · Chained tool calls · Structured output · Sustained output · Extended thinking · Vision · Counts in images · Multiple images at once · Spatial reasoning · Writes code · Reasons about code · Synthesises sources · Follows instructions
| build | publisher | precision | engine | whose work | context proved | best tok/s | machines |
|---|---|---|---|---|---|---|---|
| hf.co/unsloth/gemma-4-31B-it-GGUF:Q4_K_M | unsloth | Q4_K_M | ollama | — | 1,31,072 | 3.4 | 1 |
| hf.co/unsloth/gemma-4-31B-it-GGUF:Q4_0 | unsloth | Q4_0 | ollama | — | 1,31,072 | 2.4 | 1 |