The LLMCheck index estimates Gemma 4 26B-A4B at ~121 tok/s on an M3 Ultra at Q4_K_M, scaled from its published M5 Max run with the published bandwidth model, not a measured run on this chip. It needs about 18 GB of unified memory. It is Google's 26B mixture-of-experts model (3.8B active), released under the Apache 2.0 licence.
Want to run Gemma 4 26B-A4B faster — or step up to bigger models? Find your Mac →This is an LLMCheck index estimate from its published two-term bandwidth model, not a measured run; the index labels it so everywhere it appears. Own this Mac? Submit a real run →
| Metric | Value |
|---|---|
| Tokens per second | estimated~121 tok/s |
| Provenance | estimate · how estimates are made |
| Quantization | Q4_K_M |
| Minimum RAM | 18 GB |
| Parameters | 26B MoE (3.8B active) |
| Chip bandwidth | 819 GB/s |
Each row as the index holds it, with its source. “Mac RAM” is the memory of the Mac a run used, not the model’s requirement (18 GB). The headline figure follows the reference rule: the lower-median measured plain-decode run at 5 bits or fewer (a run whose quant was not stated counts only when no stated-quant run qualifies), else the estimate; a run marked “not used to anchor estimates” is listed here but never becomes the headline.
| Speed | Quant | Engine | Mac RAM | Context | Source | Note |
|---|---|---|---|---|---|---|
| community87.8 tok/s | MXFP4_MOE | llama.cpp | 512 GB | TG32, depth 0 | llamaperf.com | llama.cpp b8783, llama-bench, flash attention on, mean of 5, MTP off; anonymous llamaperf submission.; not used to anchor estimates |
Ollama — install it (brew install ollama or the app from ollama.com), then pull the model by its tag, checked on ollama.com on 2026-10-17:
MLX — pick a build from Hugging Face’s MLX listings, then run it with its repo id:
Downloads: Search the Ollama library → · model card →
Prefer a GUI? LM Studio downloads GGUF and MLX builds in one click. Step by step: install Ollama on a Mac.
Google's Apache 2.0 multimodal MoE: 26B total with 3.8B active (128 experts), 256K context. It fits a 24 GB Mac at 4-bit; Liquid AI's published baseline lists 107 tok/s on an M5 Max (llama.cpp Q4_K_M), and community runs measured 48–52 tok/s on a 32 GB Mac mini M6.
Apache 2.0 licence · 26B MoE · 256K context · LLMCheck capability 35/50 · see its leaderboard rank →
Gemma 4 26B-A4B needs about 18 GB of unified memory. These current Apple Silicon Macs have the headroom to run it comfortably:
As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep the LLMCheck index free.
See how Gemma 4 26B-A4B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard