Yes — Qwen3-235B-A22B (235B-A22B) runs at 19 tok/s on M3 Ultra with 256 GB RAM using Q4_K_M quantization via MLX. First token latency is 0.8s. Alibaba's massive 235B MoE (22B active) — near-cloud quality locally.
Want to run Qwen3-235B-A22B faster — or step up to bigger models? Find your Mac →The LLMCheck index estimates Qwen3-235B-A22B on M3 Ultra using our published methodology: Q4_K_M quantization, memory-bandwidth scaling, and cross-referenced third-party benchmarks where available. Figures are transparent estimates — own this config? Submit a real benchmark →
| Metric | Value |
|---|---|
| Tokens per second | 19 tok/s |
| Time to first token | 0.8s |
| Quantization | Q4_K_M |
| Minimum RAM | 256 GB |
| Recommended engine | MLX |
| Parameters | 235B-A22B |
| Benchmark date | 2026-08 |
Q4_K_M 235B-A22B MLX M3 Ultra
The recommended engine for Qwen3-235B-A22B on M3 Ultra is MLX. Install with pip and pull the model:
Alternatively, you can use Ollama for a simpler setup:
| Chip | Speed | First Token | Min RAM | Engine |
|---|---|---|---|---|
| M4 Ultra | 22 tok/s | 1.5s | 192 GB | MLX |
| M5 Max | 18 tok/s | 1.8s | 128 GB | MLX |
| M5 Max | 15 tok/s | 2.2s | 128 GB | Ollama |
To run Qwen3-235B-A22B on M3 Ultra you need:
Qwen3-235B-A22B needs about 256 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 our benchmarks free.
See how Qwen3-235B-A22B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard