Yes — Gemma 4 E2B (2.3B) runs at 82 tok/s on M3 with 16 GB RAM using Q4_K_M quantization via Ollama. First token latency is 0.3s. Google's ultra-compact 2.3B MoE model with multimodal and audio support.
LLMCheck measured Gemma 4 E2B on M3 using the standard methodology: Q4_K_M quantization, 256-token input, 512-token output, 3 runs averaged on a freshly-booted system.
| Metric | Value |
|---|---|
| Tokens per second | 82 tok/s |
| Time to first token | 0.3s |
| Quantization | Q4_K_M |
| Minimum RAM | 16 GB |
| Recommended engine | Ollama |
| Parameters | 2.3B |
| Benchmark date | 2026-04 |
Q4_K_M 2.3B Ollama M3
The recommended engine for Gemma 4 E2B on M3 is Ollama. Install Ollama, then pull the model:
Ollama handles quantization automatically — it will download the Q4_K_M variant (~16 GB) and start an interactive chat session.
| Chip | Speed | First Token | Min RAM | Engine |
|---|---|---|---|---|
| M5 Max | 158 tok/s | 0.1s | 128 GB | MLX |
| M4 Pro | 95 tok/s | 0.2s | 24 GB | Ollama |
| M1 | 58 tok/s | 0.5s | 8 GB | Ollama |
To run Gemma 4 E2B on M3 you need:
See how Gemma 4 E2B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard