MoE inference results on Mac

For systems and research readers: this is MoE inference under controlled RAM sandboxes on Apple silicon. Speed runs use the speed slider at 20 on every model and size. Exact settings keep every routed expert so you can measure fidelity and peak memory without guesswork.

Capacity and control

Peak process memory stays inside the sandbox for these runs (for example about 19.9 GB on the 80B coding MoE at 32 GB, and about 31.8 GB on the 426B-class model at 48 GB). Speed settings run 2x to 5x faster than exact on the same task: at 48 GB the 229B model goes from 7.88 to 19.2 median decode tok/s, the 235B from 2.62 to 9.19, and the 426B-class model from 2.20 to 4.62. Exact mode reports top-1 agree 1.00 across the board.

Raw results

Model RAM Mode Decode tok/s Tool call tok/s First turn s Task s Top-1 agree
Qwen3-Coder-Next 80B-A3B (6-bit)16 GBspeed41.241.817.4570.91
Qwen3-Coder-Next 80B-A3B (6-bit)16 GBexact6.446.3517.41291.00
Qwen3-Coder-Next 80B-A3B (6-bit)32 GBspeed34.935.015.9471.00
Qwen3-Coder-Next 80B-A3B (6-bit)32 GBexact13.914.816.1771.00
Qwen3-Coder-Next 80B-A3B (6-bit)48 GBspeed34.534.515.8441.00
Qwen3-Coder-Next 80B-A3B (6-bit)48 GBexact17.718.316.2571.00
Qwen3.8-Flash-Next 229B (6-bit)16 GBspeed17.116.937.31280.30
Qwen3.8-Flash-Next 229B (6-bit)16 GBexact3.533.5738.23731.00
Qwen3.8-Flash-Next 229B (6-bit)32 GBspeed18.318.131.51161.00
Qwen3.8-Flash-Next 229B (6-bit)32 GBexact6.186.0331.72331.00
Qwen3.8-Flash-Next 229B (6-bit)48 GBspeed19.219.432.21121.00
Qwen3.8-Flash-Next 229B (6-bit)48 GBexact7.887.7232.01861.00
Qwen3-235B-A22B (4-bit)16 GBspeed3.46-45.95630.19
Qwen3-235B-A22B (4-bit)16 GBexact1.131.1246.66051.00
Qwen3-235B-A22B (4-bit)32 GBspeed9.719.7144.61830.88
Qwen3-235B-A22B (4-bit)32 GBexact1.821.7046.64641.00
Qwen3-235B-A22B (4-bit)48 GBspeed9.199.2842.11490.98
Qwen3-235B-A22B (4-bit)48 GBexact2.622.3847.63431.00
MiniMax-M3 426B (4-bit)32 GBspeed3.533.3843.73660.38
MiniMax-M3 426B (4-bit)32 GBexact1.411.3944.78081.00
MiniMax-M3 426B (4-bit)48 GBspeed4.624.4744.92950.75
MiniMax-M3 426B (4-bit)48 GBexact2.202.1843.15261.00

Source: Global Agent 1.2.0 bench harness on Apple M5 Pro, 2026-10-09. Each model builds the same Express ping/pong app with a web page, installs it and starts it, using the system prompt and tools the app sends. Decode tok/s is the median across the task's requests. Speed is the speed slider at 20; exact keeps every expert the router picks. 16 and 32 GB are sandboxed on the same Mac. MiniMax-M3 needs 32 GB or more. Full interactive tables: globalagent.dev/benchmarks.

Download free for Mac - Read the docs - Licenses

Questions: info@globalagent.dev

© QWERTY Code LLC - Terms - EULA - Get started - Docs