11 builds of Gemma 4 E2B 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. 5.1B parameters, as the vendor states.
trained 128K · proved 128K context proved by unsloth IQ4_XS
174.7 tok/s peak, by unsloth IQ4_XS
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-E2B-it-GGUF:IQ4_XS | unsloth | IQ4_XS | ollama | quantizer | 1,31,072 | 174.7 | 2 |
| gemma4:e2b | Q4_K_M | ollama | vendor | 1,31,072 | 170.5 | 8 | |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:Q4_K_M | unsloth | Q4_K_M | ollama | quantizer | 1,31,072 | 164.7 | 8 |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:Q4_0 | unsloth | Q4_0 | ollama | quantizer | 1,31,072 | 84.7 | 3 |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:Q8_0 | unsloth | Q8_0 | ollama | quantizer | 1,31,072 | 67.0 | 1 |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:BF16 | unsloth | BF16 | ollama | quantizer | 1,31,072 | 38.3 | 1 |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:IQ4_NL | unsloth | IQ4_NL | ollama | quantizer | 8,192 | 54.9 | 2 |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:Q3_K_M | unsloth | Q3_K_M | ollama | quantizer | 8,192 | 38.4 | 3 |
| hf.co/unsloth/gemma-4-E2B-it-GGUF:Q3_K_S | unsloth | Q3_K_S | ollama | quantizer | 4,096 | 14.4 | 1 |
| gemma4:e2b-it-qat | ollama | — | ollama | — | — | 8.9 | 1 |
| google/gemma-4-E2B-it-qat-q4_0-gguf | — | llamacpp | — | — | 2.0 | 1 |