Llama 3.3 70B runs at 13 tok/s (13.1 as recorded) on a 128 GB M5 Max, a published figure reported by tbreak.com (MLX, 4-bit). It needs about 44 GB of unified memory. It is Meta's 70B dense model, released under the Meta Custom licence.
Want to run Llama 3.3 70B faster — or step up to bigger models? Find your Mac →This is a published figure: a real run, reported by its source and linked below. The index records it as given and never adjusts it.
| Metric | Value |
|---|---|
| Tokens per second | sourced13 tok/s |
| Provenance | published figure · tbreak.com |
| Quantization | 4-bit |
| Engine | MLX |
| Figure dated | 2026-09-21 |
| Minimum RAM | 44 GB |
| Parameters | 70B |
| Chip bandwidth | 614 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 (44 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 |
|---|---|---|---|---|---|---|
| sourced13.1 tok/s | 4-bit | MLX | 128 GB | — | tbreak.com | Mac Studio, 40-core GPU. tbreak review: 4-bit MLX, 41.2 GB peak; MLX 0.32.2 / mlx-lm 0.31.3 per the section's setup. |
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 →
The 13 tok/s figure above came from MLX at 4-bit; other runtimes and quants run at different speeds.
Prefer a GUI? LM Studio downloads GGUF and MLX builds in one click. Step by step: install Ollama on a Mac.
| Chip | Speed | Quant | Engine | Bandwidth |
|---|---|---|---|---|
| M5 Ultra | estimated23 tok/s | Q4_K_M | MLX | 1228 GB/s |
| M3 Ultra | estimated16 tok/s | Q4_K_M | Ollama | 819 GB/s |
| M1 Ultra | estimated15 tok/s | Q4_K_M | Ollama | 800 GB/s |
| M2 Max | estimated8 tok/s | Q4_K_M | Ollama | 400 GB/s |
Meta's dense 70B (Meta Custom license, 128K context). Needs a 48 GB+ Mac at 4-bit; tbreak's Mac Studio review measured 13.1 tok/s on an M5 Max (4-bit MLX, 41.2 GB peak).
Meta Custom licence · 70B · 128K context · LLMCheck capability 27/50 · see its leaderboard rank →
Llama 3.3 70B needs about 44 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 Llama 3.3 70B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard