Qwen 3.5 122B-A10B runs at 59 tok/s (58.6 as recorded) on a 512 GB M3 Ultra, a community run reported by github.com (MLX, 4-bit, 512-token prompt, 128 generated). It needs about 65 GB of unified memory. It is Alibaba's 122B mixture-of-experts model (10B active), released under the Apache 2.0 licence.
Want to run Qwen 3.5 122B-A10B faster — or step up to bigger models? Find your Mac →This is a community run: 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 | community59 tok/s |
| Provenance | community run · github.com |
| Quantization | 4-bit |
| Engine | MLX |
| Context | 512-token prompt, 128 generated |
| Figure dated | 2026-09-20 |
| Minimum RAM | 65 GB |
| Parameters | 122B MoE (10B 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 (65 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 |
|---|---|---|---|---|---|---|
| sourced60 tok/s | quantization not stated | LM Studio | 96 GB | — | bgr.com | — |
| community58.6 tok/s | 4-bit | MLX | 512 GB | 512-token prompt, 128 generated | github.com | Mac Studio, 80-core GPU, 512 GB, macOS 26.5.1. mlx_lm.benchmark -p 512 -g 128 -n 5 (mlx-lm 0.31.3, MLX 0.32.2): plain decode with no HTTP server or chat template, median of the five trial lines in the repo's engine.json; mlx-community/Qwen3.5-122B-A10B-4bit. The repo's headline test streams the same cell through mlx_lm.server (thinking and prompt cache off): 51.3 tok/s. The source discloses a stuck background process on one of the M3 Ultra's 32 CPU cores during these runs, which can only have slowed it. |
Ollama — install it (brew install ollama or the app from ollama.com), then pull the model by its tag, checked on ollama.com on 2026-10-24:
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 59 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.
Alibaba's Apache 2.0 122B MoE with 10B active (Feb 2026): Gated DeltaNet hybrid, text/image/video input, 262K context. SWE-bench Verified 72.0, GPQA Diamond 86.6, MMLU-Pro 86.7 (model card). About 65 GB at 4-bit, so it needs a 96 GB+ Mac. A community llama.cpp run measured 42.9 tok/s on an M5 Max (Q4_K_M, median of 5); BGR's review measured ~80 on an M5 Ultra and ~60 on an M3 Ultra (quantization not stated).
Apache 2.0 licence · 122B MoE · 262K context · LLMCheck capability 38/50 · see its leaderboard rank →
Qwen 3.5 122B-A10B needs about 65 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 Qwen 3.5 122B-A10B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard