Mac Studio M5 Max (128 GB) vs M5 Ultra (96 GB) for local LLMs

For most local-LLM work the verdict is the Mac Studio M5 Ultra (96 GB): a median 1.7× faster than the M5 Max (128 GB) across the 82 shared models with a figure on both; on Qwen 3.6-35B-A3B, 153 against 130 tok/s (both community runs). The M5 Max (128 GB) fits 7 more models.

Bandwidth
614 / 1,228
GB/s
Unified memory
128 / 96
GB
Models that fit
93 / 86
of 130
Qwen 3.6-35B-A3B
community130 / community153
tok/s

Spec sheet: M5 Max (128 GB) vs M5 Ultra (96 GB)

M5 Max (128 GB)M5 Ultra (96 GB)
ChipM5 MaxM5 Ultra
GPU cores40-core64- or 80-core
Memory bandwidth614 GB/s1,228 GB/s
Unified memory128 GB96 GB
Price$5,099
Apple list price, checked 14 Nov 2026 · with 512GB storage
$5,499
Apple list price, checked 14 Nov 2026 · with 1TB storage
Price per GB$40$57
Models that fit (of 130)9386
Best local LLM hereQwen 3.6-35B-A3B community130 tok/sKAT-Coder-V2.5 estimated120 tok/s
Qwen 3.6-35B-A3Bcommunity130 tok/scommunity153 tok/s
Price per tok/s on Qwen 3.6-35B-A3B$39$36
Where to buyNew · build-to-order from AppleNew · build-to-order from Apple
Bandwidth and RAM per configuration from data/macs.json (Apple’s figures); prices are Apple’s US list prices on the date shown. “Best local LLM” is the Best by Mac headline pick for each config. Winners in bold.

The models both run, side by side

Every speed below is the LLMCheck index’s figure for that exact configuration — the same number its Mac Studio M5 Max (128 GB) and Mac Studio M5 Ultra (96 GB) pages print, with its provenance mark. Across the 82 shared models with a figure on both, the M5 Ultra column is a median ×1.68 of the M5 Max column.

ModelSizeM5 Max tok/sM5 Ultra tok/sRatio
KAT-Coder-V2.535B MoEcommunity89estimated120×1.35
Qwen 3.6-35B-A3B35B MoEcommunity130community153×1.18
Ling-3.0-flash-VL124B MoEestimated53estimated89×1.68
Nanbeige4.3-33B33.1Bcommunity25estimated52×2.08
Recursive Flint39.8B MoEcommunity123estimated131×1.07
Inkling-Small caveat276B MoEestimated42estimated49×1.17
Qwen3.8-27B27.8Bcommunity31sourced48×1.55
Ling-3.0 Flash124B MoEcommunity68estimated114×1.68
Laguna S 2.1118B MoEestimated38estimated68×1.79
Nemotron 3.5 Lightning30B MoEcommunity117estimated165×1.41
North Mini Code 1.030.5B MoEcommunity120estimated150×1.25
Nemotron 3.5 Super VL 120B-A12B120B MoEcommunity36community74×2.06
Qwen 3.6-27B27Bestimated30estimated57×1.90
Kolibri-178.1B MoEestimated75estimated120×1.60
GLM-4.7-Flash31B MoEestimated107estimated140×1.31
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

Showing the top 15 of the 86 shared models, in Best-by-Mac suitability order for the Mac Studio M5 Ultra (96 GB). Every other pair of Macs: the compare tool’s two-Mac mode.

What only one of them runs

A model fits when its minimum RAM (its smallest build of at least 3 bits with a working Mac runtime, plus about 2 GB for the runtime and a short context; 2-bit only where a published Mac run used it) is at most the Mac’s unified memory. The extra memory unlocks these models from the LLMCheck index:

Only the M5 Max (128 GB): 7 more

ModelSizeMin RAMM5 Max tok/s
Qwen3.8-Flash-Next125B MoE120 GBcommunity47
DeepSeek V4 Flash284B MoE120 GBcommunity39
DeepSeek V4 Flash Vision-Exp305B MoE120 GBestimated39
GLM-5.3-Flash320B MoE128 GBestimated27
Qwen3-235B-A22B235B MoE128 GBestimated26
dots3-note preview280B MoE120 GBestimated28
Step-3.5-Flash196B MoE120 GBestimated29

Price per GB and per token/s

At Apple’s list prices the M5 Max (128 GB) costs $40 per GB of unified memory and the M5 Ultra (96 GB) $57. On Qwen 3.6-35B-A3B — the reference model here, which both run — that is $39 and $36 per tok/s of decode speed: the M5 Ultra (96 GB) buys more speed per dollar.

The $400 premium (8% more) buys +18% decode speed on Qwen 3.6-35B-A3B, but 7 fewer.

Which to buy, by use case

Bottom line: The M5 Ultra (96 GB) for speed (a median 1.7×); the M5 Max (128 GB) for 7 more models.

🛒 Where to buy these two Macs

Amazon does not stock either configuration: Apple builds both to order.

Where these numbers come from

Each speed is decode (generation) tok/s from the LLMCheck index for that exact configuration: the lower-median measured run on that chip where one exists (a filled square is a published figure, a hollow circle a community run), otherwise an estimate from the memory-bandwidth model (hollow diamond, printed with a tilde). LLMCheck runs no tests of its own. Prompt processing (prefill) depends on GPU compute rather than bandwidth, and the index does not estimate it.

FAQ: M5 Max (128 GB) vs M5 Ultra (96 GB)

Which is better for local LLMs, the Mac Studio M5 Max (128 GB) or the Mac Studio M5 Ultra (96 GB)?

For most local-LLM work the verdict is the Mac Studio M5 Ultra (96 GB): a median 1.7× faster than the M5 Max (128 GB) across the 82 shared models with a figure on both; on Qwen 3.6-35B-A3B, 153 against 130 tok/s (both community runs). The M5 Max (128 GB) fits 7 more models.

How much faster is the Mac Studio M5 Ultra (96 GB) than the Mac Studio M5 Max (128 GB) for local LLMs?

Decode speed on Apple Silicon follows memory bandwidth: the Mac Studio M5 Ultra (96 GB) reads memory at 1,228 GB/s, the Mac Studio M5 Max (128 GB) at 614 GB/s. The LLMCheck index puts Qwen 3.6-35B-A3B at 153 tok/s (community run) against 130 tok/s (community run). Across the 82 shared models with a figure on both, the median gain is 1.68x.

What can the Mac Studio M5 Max (128 GB) run that the Mac Studio M5 Ultra (96 GB) cannot?

7 models in the LLMCheck index need more than 96 GB but fit in 128 GB, led by Qwen3.8-Flash-Next, DeepSeek V4 Flash, DeepSeek V4 Flash Vision-Exp. Both run the other 86.

Is the Mac Studio M5 Ultra (96 GB) worth $400 more than the Mac Studio M5 Max (128 GB)?

It costs $5,499 against $5,099 (Apple list prices, checked 14 Nov 2026; the M5 Ultra (96 GB) price includes 1TB of storage, the M5 Max (128 GB) price 512GB). For that it gives 153 tok/s (community run) against 130 tok/s (community run) on Qwen 3.6-35B-A3B, while fitting 7 fewer. Worth it if you use the extra speed or models daily; the cheaper Mac still runs every model the dearer one does.

What is the best local LLM for each of these Macs?

On the Mac Studio M5 Max (128 GB): Qwen 3.6-35B-A3B at 130 tok/s (community run). On the Mac Studio M5 Ultra (96 GB): KAT-Coder-V2.5 at ~120 tok/s (estimated). These are the headline picks of the LLMCheck Best by Mac pages: the best-ranked model that reaches 15 tok/s on that Mac and has a standard Mac runtime.

Related