Apple M5 Max (32-core GPU), 36 GB: which models fit?

RTX 4090 RTX 5090 RTX 3090 RTX 4060 Ti (16 GB)RTX 5070 Ti Radeon RX 7900 XTX M4 Max 128 GB M3 Ultra 512 GB DGX Spark (128 GB)All devices

Apple Silicon Mac

Apple M5 Max (32-core GPU), 36 GB

Memory36 GB unifiedBandwidth460 GB/s vendor specPlanned for models27.0 GiB 75% of unified memoryFilesMLX Apple MLX format

Editorial pick · 36 GB MacsQwen3.8-27B MLX 4-bit All picks

Budget = file + KV cache + 1 GiB runtime + 0.5 GiB spare. Macs: 75% of memory. Details

32 models fit M5 Max 32c GPU · 36 GB · 8K context

Ranked by Artificial Analysis Intelligence Index, unranked models last. Arena is shown separately. How ranking works

Intelligence

Artificial Analysis Intelligence Index · top 10 models that fit M5 Max 32c GPU · 36 GB · higher is better

Qwen3.8-27B33.7
G9v3-39A5B (preview)21.8
K2-Horizon-7B20.6
Qwen3.6-35B-A3B18.2
Muse Glimmer 30B17.5
K2-Horizon-3.7B15.6
Qwen3.5-9B13.3
Qwen3.5-4B13.1
NVIDIA Nemotron 3.5 Lightning 30B-A3B12.9
IBM Granite 4.2 30B12.7

AA estimate

Generation speed

Tokens per second on M5 Max 32c GPU · 36 GB, best file for each of the same models · higher is better

Qwen3.6-35B-A3B~89.9
NVIDIA Nemotron 3.5 Lightning 30B-A3B~88.4
Qwen3.5-4B~73.7
K2-Horizon-3.7B~70.4
G9v3-39A5B (preview)~53.7
K2-Horizon-7B~39.6
Qwen3.5-9B~36.3
Qwen3.8-27B~23.6
IBM Granite 4.2 30B~23
Muse Glimmer 30B~19.5

MeasuredEstimated (~)

WeightsKV cacheRuntimeFreeMeasuredlinks to its sourceEstimatedshown with ~Unknownnever zero

Qwen3.8-27B

Qwen/Qwen3.8-27B

2026-08-1427B dense

AA index33.7rank #13

Arena1,441

Best file & memory

MLX 4-bit15.0 GiB fileFits · conditional

16.5 of 27.0 GiB+8K ctx · max ≈169K

Prompt~751 t/sEstimated

Generate~23.6 t/sEstimated

First token · 4K prompt~5.5 sEstimated

Memory budget

Weights file	14.95 GiB	mlx-community/Qwen3.8-27B-4bit · 3 files HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	16.45 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈169K tokens (native 256K).

Other MLX files

MLX 8-bit 27.5 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 15.0 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 33.7 Qwen3.8 27B (Xhigh) · Arena 1,441 · Model page

G9v3-39A5B (preview)

ai9stars/G9v3-39A5B

2026-08 (day unknown)39B · 5B active

AA index21.8 AA estimaterank #29

Arena—

Best file & memory

MLX 4-bit21.2 GiB fileFits

22.5 of 27.0 GiB8K ctx · max ≈115K

Prompt~1,470 t/sEstimated

Generate~53.7 t/sEstimated

First token · 4K prompt~2.8 sEstimated

Memory budget

Weights file	21.24 GiB	rapid-mlx/G9v3-39A5B-MLX-4bit · 5 files HF files
KV cache · 8K	0.30 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	22.53 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈115K tokens (native 128K).

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 2.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 21.8 AA estimate G9v3-39A5B · Arena unknown · Model page

K2-Horizon-7B

IFM/K2-Horizon-7B

2026-09-03size unknown

AA index20.6rank #33

Arena—

Best file & memory

MLX 8-bit8.9 GiB fileFits

11.0 of 27.0 GiB8K ctx · max ≈118K

Prompt—Unknown

Generate~39.6 t/sEstimated

First token · 4K prompt—Unknown

Memory budget

Weights file	8.91 GiB	mlx-community/K2-Horizon-7B-8bit · 2 files HF files
KV cache · 8K	1.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	11.03 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈118K tokens (native 512K).

Other MLX files

MLX 4-bit 6.7 GiB · fits · files

Where the speeds come from

Prompt unknown: no measured report, and no comparable vendor FP16 figure or device calibration.
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 8.9 GiB read per token.
First token unknown: needs a prefill value.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 20.6 K2 Horizon 7B · Arena unknown · Model page

Qwen3.6-35B-A3B

Qwen/Qwen3.6-35B-A3B

2026-04-1635B · 3B active

AA index18.2rank #36

Arena—

Best file & memory

MLX 4-bit19.0 GiB fileFits · conditional

20.2 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,450 t/sEstimated

Generate~89.9 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	19.00 GiB	mlx-community/Qwen3.6-35B-A3B-4bit · 4 files HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	20.16 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 8-bit 35.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.6 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 18.2 Qwen3.6 35B A3B (Reasoning) · Arena unknown · Model page

Muse Glimmer 30B

meta-models/Muse-Glimmer-30B

2026-08-1030B dense

AA index17.5rank #38

Arena1,389

Best file & memory

MLX 4-bit18.1 GiB fileFits

19.3 of 27.0 GiB8K ctx · max ≈128K

Prompt~685 t/sEstimated

Generate~19.5 t/sEstimated

First token · 4K prompt~6.0 sEstimated

Memory budget

Weights file	18.08 GiB	mlx-community/Muse-Glimmer-30B-4bit · 4 files HF files
KV cache · 8K	0.18 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	19.26 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other MLX files

MLX 8-bit 31.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 18.1 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 17.5 Muse Glimmer (High) · Arena 1,389 · Model page

K2-Horizon-3.7B

IFM/K2-Horizon-3.7B

2026-09-03size unknown

AA index15.6rank #41

Arena—

Best file & memory

MLX 8-bit5.0 GiB fileFits

7.1 of 27.0 GiB8K ctx · max ≈146K

Prompt—Unknown

Generate~70.4 t/sEstimated

First token · 4K prompt—Unknown

Memory budget

Weights file	5.01 GiB	mlx-community/K2-Horizon-3.7B-8bit · 2 files HF files
KV cache · 8K	1.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	7.13 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈146K tokens (native 512K).

Other MLX files

MLX 4-bit 3.9 GiB · fits · files

Where the speeds come from

Prompt unknown: no measured report, and no comparable vendor FP16 figure or device calibration.
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 5.0 GiB read per token.
First token unknown: needs a prefill value.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 15.6 K2 Horizon 3.7B · Arena unknown · Model page

Qwen3.5-9B

Qwen/Qwen3.5-9B

2026-03-029B dense

AA index13.3 AA estimaterank #45

Arena—

Best file & memory

MLX 8-bit9.7 GiB fileFits · conditional

11.0 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,252 t/sEstimated

Generate~36.3 t/sEstimated

First token · 4K prompt~1.8 sEstimated

Memory budget

Weights file	9.71 GiB	mlx-community/Qwen3.5-9B-8bit · 2 files HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	10.96 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 5.5 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 9.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 13.3 AA estimate Qwen3.5 9B (Non-reasoning) · Arena unknown · Model page

Qwen3.5-4B

Qwen/Qwen3.5-4B

2026-03-024B dense

AA index13.1 AA estimaterank #48

Arena—

Best file & memory

MLX 8-bit4.8 GiB fileFits · conditional

6.0 of 27.0 GiB+8K ctx · max ≈256K

Prompt~5,068 t/sEstimated

Generate~73.7 t/sEstimated

First token · 4K prompt~808 msEstimated

Memory budget

Weights file	4.78 GiB	mlx-community/Qwen3.5-4B-8bit HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	6.03 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 2.8 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 4.8 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 13.1 AA estimate Qwen3.5 4B (Reasoning) · Arena unknown · Model page

NVIDIA Nemotron 3.5 Lightning 30B-A3B

nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16

2026-08-1130B · 3B active

AA index12.9rank #49

Arena1,329

Best file & memory

MLX 4-bit16.6 GiB fileFits · conditional

17.6 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,450 t/sEstimated

Generate~88.4 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	16.55 GiB	mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-4bit · 4 files HF files
KV cache · 8K	0.05 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	17.60 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 8-bit 31.3 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 12.9 Nemotron 3.5 Lightning · Arena 1,329 · Model page

IBM Granite 4.2 30B

ibm-granite/granite-4.2-30b

2026-08-2530B dense

AA index12.7rank #50

Arena1,361

Best file & memory

MLX 4-bit15.3 GiB fileFits

18.3 of 27.0 GiB8K ctx · max ≈41K

Prompt~676 t/sEstimated

Generate~23 t/sEstimated

First token · 4K prompt~6.1 sEstimated

Memory budget

Weights file	15.34 GiB	lmstudio-community/granite-4.2-30b-MLX-4bit · 4 files HF files
KV cache · 8K	2.00 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	18.34 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈41K tokens (native 128K).

Other MLX files

MLX 8-bit 29.0 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 15.3 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 12.7 Granite 4.2 30B · Arena 1,361 · Model page

MiniCPM5-2B

openbmb/MiniCPM5-2B

2026-09-072.5B dense

AA index12.5rank #51

Arena—

Best file & memory

MLX 8-bit2.5 GiB fileFits

3.8 of 27.0 GiB8K ctx · max ≈128K

Prompt~8,055 t/sEstimated

Generate~142 t/sEstimated

First token · 4K prompt~509 msEstimated

Memory budget

Weights file	2.49 GiB	mlx-community/MiniCPM5-2B-8bit HF files
KV cache · 8K	0.33 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	3.82 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other MLX files

MLX 4-bit 1.3 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 2.5 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 12.5 MiniCPM5-2B · Arena unknown · Model page

Nemotron Cascade 2 30B A3B

nvidia/Nemotron-Cascade-2-30B-A3B

2026-03-1930B · 3B active

AA index11.7 AA estimaterank #53

Arena—

Best file & memory

MLX 4-bit16.6 GiB fileFits · conditional

17.6 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,450 t/sEstimated

Generate~88.4 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	16.55 GiB	mlx-community/Nemotron-Cascade-2-30B-A3B-4bit · 4 files HF files
KV cache · 8K	0.05 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	17.60 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 8-bit 31.3 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 11.7 AA estimate Nemotron Cascade 2 30B A3B · Arena unknown · Model page

Granite 4.2 8B

ibm-granite/granite-4.2-8b

2026-08-258B dense

AA index11.1rank #56

Arena1,320

Best file & memory

MLX 8-bit8.7 GiB fileFits

10.9 of 27.0 GiB8K ctx · max ≈108K

Prompt~2,534 t/sEstimated

Generate~40.5 t/sEstimated

First token · 4K prompt~1.6 sEstimated

Memory budget

Weights file	8.70 GiB	lmstudio-community/granite-4.2-8b-MLX-8bit · 2 files HF files
KV cache · 8K	1.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	10.95 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈108K tokens (native 128K).

Other MLX files

MLX 4-bit 4.6 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 8.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 11.1 Granite 4.2 8B · Arena 1,320 · Model page

G9v3-3B

AI9Stars/G9v3-3B

2026-07 (day unknown)3B dense

AA index10.8 AA estimaterank #58

Arena—

Best file & memory

MLX 4-bit1.6 GiB fileFits

3.0 of 27.0 GiB8K ctx · max ≈128K

Prompt~6,757 t/sEstimated

Generate~225 t/sEstimated

First token · 4K prompt~606 msEstimated

Memory budget

Weights file	1.57 GiB	cof139/G9v3-3B-mlx-4Bit HF files
KV cache · 8K	0.41 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	2.97 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 1.6 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 10.8 AA estimate G9v3-3B · Arena unknown · Model page

North Mini Code

CohereLabs/North-Mini-Code-1.0

2026-06-0930B · 3B active

AA index9.9rank #61

Arena—

Best file & memory

MLX 4-bit17.2 GiB fileFits

18.7 of 27.0 GiB8K ctx · max ≈250K

Prompt~2,450 t/sEstimated

Generate~85 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	17.23 GiB	mlx-community/North-Mini-Code-1.0-4bit · 4 files HF files
KV cache · 8K	0.48 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	18.71 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈250K tokens (native 250K).

Other MLX files

MLX 8-bit 31.0 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.9 North Mini Code · Arena unknown · Model page

Granite 4.2 3B

ibm-granite/granite-4.2-3b

2026-08-253B dense

AA index9.1rank #64

Arena1,293

Best file & memory

MLX 8-bit3.6 GiB fileFits

5.2 of 27.0 GiB8K ctx · max ≈128K

Prompt~6,757 t/sEstimated

Generate~97.4 t/sEstimated

First token · 4K prompt~606 msEstimated

Memory budget

Weights file	3.62 GiB	lmstudio-community/granite-4.2-3b-MLX-8bit HF files
KV cache · 8K	0.63 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	5.25 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other MLX files

MLX 4-bit 1.9 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 3.6 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.1 Granite 4.2 3B · Arena 1,293 · Model page

MiniCPM5-1B

openbmb/MiniCPM5-1B

2026-05-191.1B dense

AA index8.8 AA estimaterank #67

Arena—

Best file & memory

MLX 8-bit1.1 GiB fileFits

2.3 of 27.0 GiB8K ctx · max ≈128K

Prompt~18,759 t/sEstimated

Generate~330 t/sEstimated

First token · 4K prompt~218 msEstimated

Memory budget

Weights file	1.07 GiB	mlx-community/MiniCPM5-1B-8bit HF files
KV cache · 8K	0.19 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	2.26 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other MLX files

MLX 4-bit 0.6 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 1.1 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.8 AA estimate MiniCPM5-1B (Reasoning) · Arena unknown · Model page

Nanbeige4.1-3B

Nanbeige/Nanbeige4.1-3B

2026-02 (day unknown)3B dense

AA index8.4 AA estimaterank #69

Arena—

Best file & memory

MLX 8-bit3.9 GiB fileFits

5.4 of 27.0 GiB8K ctx · max ≈256K

Prompt~6,757 t/sEstimated

Generate~90.6 t/sEstimated

First token · 4K prompt~606 msEstimated

Memory budget

Weights file	3.89 GiB	mlx-community/Nanbeige4.1-3B-8bit HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	5.39 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 2.1 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 3.9 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.4 AA estimate Nanbeige4.1-3B · Arena unknown · Model page

LFM2.5-2.6B

LiquidAI/LFM2.5-2.6B

2026-08-042.7B dense

AA index8.4 AA estimaterank #70

Arena—

Best file & memory

MLX 8-bit2.7 GiB fileFits · conditional

3.8 of 27.0 GiB+8K ctx · max ≈128K

Prompt~7,536 t/sEstimated

Generate~132 t/sEstimated

First token · 4K prompt~544 msEstimated

Memory budget

Weights file	2.67 GiB	LiquidAI/LFM2.5-2.6B-MLX-8bit HF files
KV cache · 8K	0.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	3.79 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other MLX files

MLX 4-bit 1.5 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 2.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.4 AA estimate LFM2.5-2.6B · Arena unknown · Model page

Falcon-H1R-7B

tiiuae/Falcon-H1R-7B

2026-01-057B dense

AA index7.8 AA estimaterank #71

Arena—

Best file & memory

MLX 8-bit7.5 GiB fileFits · conditional

8.9 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,896 t/sEstimated

Generate~47 t/sEstimated

First token · 4K prompt~1.4 sEstimated

Memory budget

Weights file	7.51 GiB	mlx-community/Falcon-H1R-7B-8bit · 2 files HF files
KV cache · 8K	0.34 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	8.85 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 4.0 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 7.5 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.8 AA estimate Falcon-H1R-7B · Arena unknown · Model page

LFM2.5-8B-A1B

LiquidAI/LFM2.5-8B-A1B

2026-05-288.3B · 1.5B active

AA index7.2 AA estimaterank #74

Arena—

Best file & memory

MLX 8-bit8.4 GiB fileFits · conditional

9.5 of 27.0 GiB+8K ctx · max ≈125K

Prompt~4,900 t/sEstimated

Generate~96.6 t/sEstimated

First token · 4K prompt~836 msEstimated

Memory budget

Weights file	8.38 GiB	LiquidAI/LFM2.5-8B-A1B-MLX-8bit · 2 files HF files
KV cache · 8K	0.09 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	9.48 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈125K tokens (native 125K).

Other MLX files

MLX 4-bit 4.5 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.5 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.2 AA estimate LFM2.5-8B-A1B · Arena unknown · Model page

NVIDIA Nemotron 3 Nano 4B

nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16

2026-03-164B dense

AA index4.8rank #75

Arena—

Best file & memory

MLX 8-bit3.9 GiB fileFits · conditional

5.1 of 27.0 GiB+8K ctx · max ≈256K

Prompt~5,106 t/sEstimated

Generate~89.7 t/sEstimated

First token · 4K prompt~802 msEstimated

Memory budget

Weights file	3.93 GiB	mlx-community/NVIDIA-Nemotron-3-Nano-4B-8bit HF files
KV cache · 8K	0.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	5.06 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 2.1 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 3.9 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 4.8 NVIDIA Nemotron 3 Nano 4B · Arena unknown · Model page

LFM2.5-24B-A2B

LiquidAI/LFM2.5-24B-A2B

2026-10-1324B · 2B active

AA index—unranked

Arena—

Best file & memory

MLX 8-bit23.7 GiB fileFits · conditional

24.9 of 27.0 GiB+8K ctx · max ≈90K

Prompt~3,675 t/sEstimated

Generate~74 t/sEstimated

First token · 4K prompt~1.1 sEstimated

Memory budget

Weights file	23.73 GiB	LiquidAI/LFM2.5-24B-A2B-MLX-8bit · 5 files HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	24.89 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈90K tokens (native 128K).

Other MLX files

MLX 4-bit 12.6 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 2.0 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

MiniCPM-V 4.7 35B-A3B

openbmb/MiniCPM-V-4.7-35B-A3B

2026-10-1235B · 3B active

AA index—unranked

Arena—

Best file & memory

MLX 4-bit21.9 GiB fileFits · conditional

23.1 of 27.0 GiB+8K ctx · max ≈184K

Prompt~2,450 t/sEstimated

Generate~77.9 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	21.91 GiB	mlx-community/MiniCPM-V-4.7-35B-A3B-4bit · 5 files HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	23.07 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈184K tokens (native 256K).

Other MLX files

MLX 8-bit 37.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.9 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

MiMo-V2.6-Distill-Qwen-9B

XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B

2026-09-229B dense

AA index—unranked

Arena—

Best file & memory

MLX 8-bit9.7 GiB fileFits · conditional

11.0 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,252 t/sEstimated

Generate~36.3 t/sEstimated

First token · 4K prompt~1.8 sEstimated

Memory budget

Weights file	9.71 GiB	nativ-community/MiMo-V2.6-Distill-Qwen-9B-MLX-8bit · 2 files HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	10.96 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 5.5 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 9.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

K2-Horizon-0.9B

IFM/K2-Horizon-0.9B

2026-09-03size unknown

AA index—unranked

Arena—

Best file & memory

MLX 8-bit1.1 GiB fileFits

2.5 of 27.0 GiB8K ctx · max ≈128K

Prompt—Unknown

Generate~330 t/sEstimated

First token · 4K prompt—Unknown

Memory budget

Weights file	1.07 GiB	mlx-community/K2-Horizon-0.9B-8bit HF files
KV cache · 8K	0.44 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	2.50 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other MLX files

MLX 4-bit 0.8 GiB · fits · files

Where the speeds come from

Prompt unknown: no measured report, and no comparable vendor FP16 figure or device calibration.
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 1.1 GiB read per token.
First token unknown: needs a prefill value.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Ornith-1.5-9B

ornith-ai/Ornith-1.5-9B

2026-08 (day unknown)9B dense

AA index—unranked

Arena—

Best file & memory

MLX 8-bit8.9 GiB fileFits · conditional

10.1 of 27.0 GiB+8K ctx · max ≈256K

Prompt~2,252 t/sEstimated

Generate~39.8 t/sEstimated

First token · 4K prompt~1.8 sEstimated

Memory budget

Weights file	8.86 GiB	ornith-ai/Ornith-1.5-9B-MLX-8bit · 2 files HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	10.11 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 4.7 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 8.9 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.6-27B

Qwen/Qwen3.6-27B

2026-04-2227B dense

AA index—unranked

Arena—

Best file & memory

MLX 4-bit15.0 GiB fileFits · conditional

16.5 of 27.0 GiB+8K ctx · max ≈169K

Prompt750 t/sMeasured

Generate24.8 t/sMeasured

First token · 4K prompt~5.5 sEstimated

Memory budget

Weights file	14.95 GiB	mlx-community/Qwen3.6-27B-4bit · 3 files HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	16.45 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈169K tokens (native 256K).

Other MLX files

MLX 8-bit 27.5 GiB · too big · files

Where the speeds come from

Prompt measured 750 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-09-17 · omlx.ai. Reported by the source; build and settings may differ from yours.
Generate measured 24.8 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-09-17 · omlx.ai. Reported by the source; build and settings may differ from yours.
First token 4K-token prompt ÷ measured prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-0.8B

Qwen/Qwen3.5-0.8B

2026-03-020.8B dense

AA index—unranked

Arena—

Best file & memory

MLX 8-bit0.9 GiB fileFits · conditional

2.0 of 27.0 GiB+8K ctx · max ≈256K

Prompt~25,340 t/sEstimated

Generate~378 t/sEstimated

First token · 4K prompt~162 msEstimated

Memory budget

Weights file	0.93 GiB	mlx-community/Qwen3.5-0.8B-8bit HF files
KV cache · 8K	0.09 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	2.03 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 0.6 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 0.9 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-2B

Qwen/Qwen3.5-2B

2026-03-022B dense

AA index—unranked

Arena—

Best file & memory

MLX 8-bit2.5 GiB fileFits · conditional

3.6 of 27.0 GiB+8K ctx · max ≈256K

Prompt~10,136 t/sEstimated

Generate~142 t/sEstimated

First token · 4K prompt~404 msEstimated

Memory budget

Weights file	2.48 GiB	mlx-community/Qwen3.5-2B-8bit HF files
KV cache · 8K	0.09 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	3.57 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 4-bit 1.6 GiB · fits · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 2.5 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-27B

Qwen/Qwen3.5-27B

2026-02-2427B dense

AA index—unranked

Arena—

Best file & memory

MLX 4-bit15.0 GiB fileFits · conditional

16.5 of 27.0 GiB+8K ctx · max ≈169K

Prompt~751 t/sEstimated

Generate~23.6 t/sEstimated

First token · 4K prompt~5.5 sEstimated

Memory budget

Weights file	14.95 GiB	mlx-community/Qwen3.5-27B-4bit · 3 files HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	16.45 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈169K tokens (native 256K).

Other MLX files

MLX 8-bit 27.5 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 15.0 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-35B-A3B

Qwen/Qwen3.5-35B-A3B

2026-02-2435B · 3B active

AA index—unranked

Arena—

Best file & memory

MLX 4-bit19.0 GiB fileFits · conditional

20.1 of 27.0 GiB+8K ctx · max ≈256K

Prompt3,548 t/sMeasured

Generate118 t/sMeasured

First token · 4K prompt~1.2 sEstimated

Memory budget

Weights file	18.99 GiB	mlx-community/Qwen3.5-35B-A3B-4bit · 4 files HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	20.15 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other MLX files

MLX 8-bit 35.1 GiB · too big · files

Where the speeds come from

Prompt measured 3,548 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-08-03 · omlx.ai. Reported by the source; build and settings may differ from yours.
Generate measured 118 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-08-03 · omlx.ai. Reported by the source; build and settings may differ from yours.
First token 4K-token prompt ÷ measured prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

May fit: KV-cache size unknown 10

Weights and runtime fit, but the cache for this attention type is not established. Ranked by AA index.

K2-Horizon-MoVA-36B-A4B

IFM/K2-Horizon-MoVA-36B-A4B

2026-09-0336B · 4B active

AA index25.3rank #22

Arena—

Best file & memory

MLX 4-bit24.6 GiB fileCache unknown

25.6 of 27.0 GiB+8K ctx

Prompt~1,838 t/sEstimated

Generate~53.6 t/sEstimated

First token · 4K prompt~2.2 sEstimated

Memory budget

Weights file	24.59 GiB	mlx-community/K2-Horizon-MoVA-36B-A4B-4bit · 5 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	25.59 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 512K).

Other MLX files

MLX 8-bit 37.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 2.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 25.3 K2 Horizon MoVA 36B A4B · Arena unknown · Model page

Gemma 4 26B A4B (it)

google/gemma-4-26B-A4B-it

2026-04-0225B · 3.8B active

AA index16.7 AA estimaterank #39

Arena1,434

Best file & memory

MLX 4-bit14.3 GiB fileCache unknown

15.3 of 27.0 GiB+8K ctx

Prompt~1,934 t/sEstimated

Generate~67.9 t/sEstimated

First token · 4K prompt~2.1 sEstimated

Memory budget

Weights file	14.29 GiB	mlx-community/gemma-4-26B-A4B-it-4bit · 3 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	15.29 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other MLX files

MLX 8-bit 26.0 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 2.2 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 16.7 AA estimate Gemma 4 26B A4B (Reasoning) · Arena 1,434 · Model page

Gemma 4 31B (it)

google/gemma-4-31B-it

2026-04-0231B dense

AA index14.7rank #42

Arena1,443

Best file & memory

MLX 4-bit17.1 GiB fileCache unknown

18.1 of 27.0 GiB+8K ctx

Prompt590 t/sMeasured

Generate21.3 t/sMeasured

First token · 4K prompt~6.9 sEstimated

Memory budget

Weights file	17.15 GiB	mlx-community/gemma-4-31B-it-4bit · 4 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	18.15 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other MLX files

MLX 8-bit 31.4 GiB · too big · files

Where the speeds come from

Prompt measured 590 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-10-13 · omlx.ai. Reported by the source; build and settings may differ from yours.
Generate measured 21.3 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-10-13 · omlx.ai. Reported by the source; build and settings may differ from yours.
First token 4K-token prompt ÷ measured prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 14.7 Gemma 4 31B (Reasoning) · Arena 1,443 · Model page

Gemma 4 12B (it)

google/gemma-4-12B-it

2026-06-0312B dense

AA index14.2 AA estimaterank #44

Arena—

Best file & memory

MLX 8-bit11.8 GiB fileCache unknown

12.8 of 27.0 GiB+8K ctx

Prompt~1,696 t/sEstimated

Generate~29.8 t/sEstimated

First token · 4K prompt~2.4 sEstimated

Memory budget

Weights file	11.84 GiB	mlx-community/gemma-4-12B-it-8bit · 3 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	12.84 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other MLX files

MLX 4-bit 6.3 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 11.8 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 14.2 AA estimate Gemma 4 12B (Reasoning) · Arena unknown · Model page

Ling-3.0-tiny

inclusionAI/Ling-3.0-tiny

2026-08 (day unknown)7.9B · 1.3B active

AA index11.1rank #57

Arena—

Best file & memory

MLX 4-bit4.1 GiB fileCache unknown

5.1 of 27.0 GiB+8K ctx

Prompt~5,654 t/sEstimated

Generate~215 t/sEstimated

First token · 4K prompt~724 msEstimated

Memory budget

Weights file	4.14 GiB	rapid-mlx/Ling-3.0-tiny-MLX-4bit HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	5.14 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 128K).

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 0.7 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 11.1 Ling 3.0 Tiny · Arena unknown · Model page

Nemotron 3 Nano Omni 30B A3B Reasoning

nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16

2026-04-2830B · 3B active

AA index10.3 AA estimaterank #60

Arena—

Best file & memory

MLX 4-bit18.3 GiB fileCache unknown

19.3 of 27.0 GiB+8K ctx

Prompt~2,450 t/sEstimated

Generate~80 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	18.29 GiB	mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-4bit · 4 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	19.29 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other MLX files

MLX 8-bit 33.3 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 1.8 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 10.3 AA estimate Nemotron 3 Nano Omni 30B A3B Reasoning · Arena unknown · Model page

DiffusionGemma 26B A4B

google/diffusiongemma-26B-A4B-it

2026-06-1025B · 3.8B active

AA index9.5 AA estimaterank #62

Arena—

Best file & memory

MLX 4-bit15.4 GiB fileCache unknown

16.4 of 27.0 GiB+8K ctx

Prompt~1,934 t/sEstimated

Generate~63 t/sEstimated

First token · 4K prompt~2.1 sEstimated

Memory budget

Weights file	15.41 GiB	mlx-community/diffusiongemma-26B-A4B-it-4bit · 4 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	16.41 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other MLX files

MLX 8-bit 26.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 37 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.3416 (n=94) × memory bandwidth ÷ 2.3 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.5 AA estimate DiffusionGemma 26B A4B · Arena unknown · Model page

Gemma 4 E4B (it)

google/gemma-4-E4B-it

2026-04-028B dense

AA index8.9rank #66

Arena—

Best file & memory

MLX 8-bit8.3 GiB fileCache unknown

9.3 of 27.0 GiB+8K ctx

Prompt~2,534 t/sEstimated

Generate~42.6 t/sEstimated

First token · 4K prompt~1.6 sEstimated

Memory budget

Weights file	8.27 GiB	mlx-community/gemma-4-E4B-it-8bit · 2 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	9.27 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 128K).

Other MLX files

MLX 4-bit 4.8 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 8.3 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.9 Gemma 4 E4B (Reasoning) · Arena unknown · Model page

Gemma 4 E2B

google/gemma-4-E2B-it

2026-04-025.1B dense

AA index7.8 AA estimaterank #72

Arena—

Best file & memory

MLX 8-bit5.5 GiB fileCache unknown

6.5 of 27.0 GiB+8K ctx

Prompt~3,975 t/sEstimated

Generate~64.5 t/sEstimated

First token · 4K prompt~1.0 sEstimated

Memory budget

Weights file	5.46 GiB	mlx-community/gemma-4-E2B-it-8bit · 2 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	6.46 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 128K).

Other MLX files

MLX 4-bit 3.3 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median prefill rate per GPU core of 16 measured M5 points (MLX) × this chip's GPU cores ÷ (2 × active params).
Generate estimated from MLX median efficiency 0.8231 (n=60) × memory bandwidth ÷ 5.5 GiB read per token.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.8 AA estimate Gemma 4 E2B (Reasoning) · Arena unknown · Model page

GLM-4.7-Flash

zai-org/GLM-4.7-Flash

2026-01 (day unknown)30B · 3B active

AA index—unranked

Arena—

Best file & memory

MLX 4-bit15.7 GiB fileCache unknown

16.7 of 27.0 GiB+8K ctx

Prompt1,696 t/sMeasured

Generate71.6 t/sMeasured

First token · 4K prompt~2.4 sEstimated

Memory budget

Weights file	15.70 GiB	mlx-community/GLM-4.7-Flash-4bit · 4 files HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	16.70 GiB	of 27.0 GiB usable (75% of 36 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 198K).

Other MLX files

MLX 8-bit 29.6 GiB · too big · files

Where the speeds come from

Prompt measured 1,696 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-09-28 · omlx.ai. Reported by the source; build and settings may differ from yours.
Generate measured 71.6 t/s: MLX · quant label MLX 4bit · oMLX benchmark, 4096-token prompt (PP) and generation (TG), no MTP/DFlash/SpecPrefill · 2026-09-28 · omlx.ai. Reported by the source; build and settings may differ from yours.
First token 4K-token prompt ÷ measured prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

29 models too big for M5 Max 32c GPU · 36 GB

GLM-5.3-FlashMLX 4-bit needs 191 GiB+ (KV unknown)
Qwen3.8-Flash-NextMLX 4-bit needs 105 GiB+ (KV unknown)
Muse Spark Open (816B-A42B)MLX 4-bit needs 498 GiB+ (KV unknown)
Motif 3MLX 8-bit needs 313 GiB+ (KV unknown)
MiniMax-M3MLX 4-bit needs 226 GiB+ (KV unknown)
Nex-N2-Pro (Qwen3.5-397B-A17B based)MLX 8-bit needs 394 GiB
NVIDIA Nemotron 3.5 Super VL 120B-A12BMLX 4-bit needs 66.2 GiB
Kimi K2.7 CodeMLX 4-bit needs 598 GiB+ (KV unknown)
Inkling-SmallMLX 4-bit needs 145 GiB
InklingMLX 4-bit needs 513 GiB
Solar Open 2 (250B-A15B)MLX 4-bit needs 133 GiB
NVIDIA Nemotron 3 Ultra 550B-A55BMLX 4-bit needs 290 GiB
Qwen3.5-397B-A17BMLX 4-bit needs 210 GiB
NVIDIA Nemotron 3.5 Bolt 74B-A9BMLX 4-bit needs 41.8 GiB
Qwen3.5-122B-A10BMLX 4-bit needs 66.5 GiB
Mistral Medium 3.5 (128B dense)MLX 4-bit needs 76.9 GiB
LongCat-Flash-LiteMLX 4-bit needs 37.4 GiB+ (KV unknown)
Trinity-Large-ThinkingMLX 4-bit needs 212 GiB
Qwen3-Coder-NextMLX 4-bit needs 43.5 GiB
Aleph Alpha Kolibri-1MLX 4-bit needs 43.0 GiB
Hy4-previewMLX 4-bit needs 405 GiB+ (KV unknown)
Ling-3.0-flashMLX 4-bit needs 66.7 GiB+ (KV unknown)
DeepSeek V4 Flash 0731MLX 4-bit needs 145 GiB+ (KV unknown)
Poolside Laguna S 2.1MLX 4-bit needs 63.6 GiB
GLM-5.2MLX 4-bit needs 391 GiB+ (KV unknown)
Step 3.7 FlashMLX 4-bit needs 105 GiB+ (KV unknown)
Kimi K2.6MLX 4-bit needs 614 GiB+ (KV unknown)
NVIDIA Nemotron 3 Super 120B-A12BMLX 4-bit needs 64.9 GiB
MiniMax-M2.7MLX 4-bit needs 123 GiB

Specs & sources

Memory	36 GB unified	source
Bandwidth	460 GB/s	source
FP16	unknown	Apple publishes no comparable FP16 throughput or GPU-only TDP in reviewed chip/device technical specifications. Power adapter rating and whole-system maximum continuous power are not GPU TDP.
Power	unknown	Apple publishes no comparable FP16 throughput or GPU-only TDP in reviewed chip/device technical specifications. Power adapter rating and whole-system maximum continuous power are not GPU TDP.

Found in: Mac Studio (M5 Max or M5 Ultra)

Measured speeds

4 reported observations

On this device (same chip and GPU-core count). Reported by the linked sources; not reproduced by LocalFit.

Open in speed database Submit yours

Doesn't fit? Rent a GPU

RunpodPlaceholderCloud GPU pods you can rent for running larger models.
Vast.aiPlaceholderGPU rental marketplace with many hosts and GPU types.
LambdaPlaceholderCloud GPU instances for AI training and inference.
TensorDockPlaceholderGPU cloud marketplace with consumer and datacenter GPUs.

Sponsored placeholder links: LocalFit may earn a commission if these become affiliate links. They never influence fit results, estimates or picks.