VOOZH about

URL: https://willitrunai.com/my-results/m4-max-96gb

⇱ MacBook Pro M4 Max 96GB AI Tier List: 329 models ranked | WillItRunAI


AI Compatibility Tier List

MacBook Pro M4 Max 96GB

374 AI models ranked by how well they run on your 96 GB unified memory hardware. Best match: Qwen 3.6 35B A3B (score: 93).

374

Models tested

329

Can run

96 GB

unified memory

44

Best tok/s

Tier Distribution

S 22
A 60
B 54
C 193
F 45
S: 22A: 60B: 54C: 193F: 45
S

S Tier -- Perfect Fit (22 models)

These models run flawlessly with headroom to spare.

👁 Alibaba
Qwen 3.6 35B A3B
S93
35B36.7 GB44 tok/s142K ctx
+1moe
👁 Alibaba
Qwen3-Coder 30B A3B Instruct
S93
30.5B31.3 GB52 tok/s256K ctx
moe
👁 Alibaba
Qwen3-VL 30B A3B Instruct
S92
30B31.0 GB54 tok/s256K ctx
moe
👁 Alibaba
Qwen 3.5 35B A3B
S92
35B34.1 GB48 tok/s131K ctx
moe
👁 Alibaba
Qwen 3.5 27B
S91
27B30.9 GB36 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 32B
S90
32B34.7 GB34 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 30B A3B
S90
30.5B31.3 GB52 tok/s131K ctx
moe
👁 Alibaba
Qwen 3.6 27B
S89
27B28.7 GB27 tok/s262K ctx
+1dense
👁 Alibaba
Qwen3-Coder-Next
S89
80B61.5 GB23 tok/s99K ctx
moe
👁 Mistral
Magistral Small 2507
S89
24B28.3 GB37 tok/s131K ctx
dense
👁 NVIDIA
Nemotron 3 Nano 30B
S89
30B32.0 GB38 tok/s131K ctx
dense
👁 Mistral
Devstral Small 2 24B Instruct
S89
24B28.3 GB37 tok/s256K ctx
dense
👁 Alibaba
Qwen 3.5 9B
S89
9B19.0 GB73 tok/s131K ctx
dense
👁 Alibaba
Qwen 2.5 VL 72B
S88
72B60.1 GB15 tok/s33K ctx
dense
👁 Google
Gemma 4 31B
S88
30.7B44.6 GB26 tok/s43K ctx
dense
👁 NVIDIA
Nemotron Cascade 2 30B A3B
S88
30B32.5 GB53 tok/s216K ctx
moe
👁 OpenAI
GPT-OSS 20B
S88
21B26.5 GB66 tok/s128K ctx
moe
👁 Mistral
Devstral Small 1.1
S87
24B28.3 GB37 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 8B
S87
8B18.3 GB83 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 14B
S87
14B22.2 GB38 tok/s131K ctx
dense
👁 Microsoft
Phi-4-reasoning-plus 14B
S86
14.7B23.3 GB38 tok/s33K ctx
dense
👁 Alibaba
Qwen 3.5 4B
S85
4B15.9 GB56 tok/s131K ctx
dense
A

A Tier -- Great Fit (60 models)

Excellent performance with comfortable memory margins.

👁 Google
Gemma 4 26B A4B
A85
25.2B30.3 GB56 tok/s186K ctx
moe
👁 LG AI
EXAONE 4.0 32B
A84
32B34.7 GB33 tok/s131K ctx
dense
👁 Alibaba
Qwen 2.5 32B
A83
32B34.7 GB33 tok/s131K ctx
dense
👁 NVIDIA
Nemotron Nano 8B
A82
8B18.1 GB89 tok/s131K ctx
dense
👁 Google
Gemma 3 27B
A82
27B39.0 GB27 tok/s59K ctx
dense
👁 Meta
Llama 3.3 70B
A82
70B58.9 GB15 tok/s50K ctx
dense
👁 Moonshot AI
Kimi Linear 48B A3B
A82
48B41.5 GB21 tok/s493K ctx
linear
👁 Microsoft
Phi-4 Mini Reasoning 4B
A82
3.8B15.1 GB53 tok/s131K ctx
dense
👁 Allen AI
OLMo 2 32B
A82
32B34.7 GB33 tok/s4K ctx
dense
👁 Mistral
Mistral Small 3.2 24B
A82
24B28.3 GB37 tok/s131K ctx
vision
👁 IBM
Granite 4.1 30B
A82
30B33.5 GB35 tok/s131K ctx
dense
👁 Mistral
Codestral 2 25.08
A81
22B27.1 GB35 tok/s256K ctx
dense
👁 Mistral
Ministral 3 14B
A81
14B22.2 GB38 tok/s262K ctx
multimodal
A81
24B28.3 GB37 tok/s131K ctx
dense
A80
19B25.3 GB39 tok/s8K ctx
dense
👁 Mistral
Mistral Small 24B
A80
24B28.3 GB37 tok/s33K ctx
dense
👁 InternLM
InternVL2 8B
A79
8B18.1 GB83 tok/s8K ctx
dense
👁 Meta
Llama 3.1 70B
A79
70B58.9 GB15 tok/s50K ctx
dense
👁 Alibaba
Qwen 2.5 72B
A79
72B60.1 GB15 tok/s46K ctx
dense
👁 Mistral AI
Mistral Small 3.1 24B
A79
24B28.3 GB37 tok/s131K ctx
dense
A78
14B22.9 GB38 tok/s16K ctx
dense
👁 Mistral
Ministral 3 8B
A78
8B18.3 GB83 tok/s262K ctx
multimodal
👁 DeepSeek
DeepSeek Coder V2 16B
A78
16B24.3 GB84 tok/s131K ctx
moe
👁 OpenBMB
MiniCPM-V 2.6 8B
A78
8B18.1 GB83 tok/s2K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 32B
A77
32B34.7 GB33 tok/s131K ctx
dense
👁 IBM
Granite Code 20B
A77
20B26.6 GB38 tok/s8K ctx
dense
👁 Defog
SQLCoder 7B
A77
7B17.5 GB94 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 14B
A77
14B22.7 GB38 tok/s131K ctx
dense
👁 NVIDIA
Nemotron Nano 9B v2
A77
9B19.2 GB79 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 4B
A77
4B15.9 GB56 tok/s33K ctx
dense
👁 Alibaba
Qwen 2.5 VL 7B
A77
7B16.4 GB95 tok/s33K ctx
dense
👁 IBM
Granite Code 34B
A76
34B35.7 GB31 tok/s8K ctx
dense
👁 Google
Gemma 3 12B
A76
12B23.5 GB33 tok/s131K ctx
dense
👁 Cohere
Command R 35B
A76
35B35.1 GB31 tok/s131K ctx
dense
👁 Jina AI
Jina Embeddings v3
A75
0.57B14.4 GB8 tok/s8K ctx
dense
👁 Magistral AI
Magistral 7B
A75
7B17.5 GB94 tok/s8K ctx
dense
👁 BigCode
StarCoder 15B
A75
15B36.7 GB30 tok/s8K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 32B
A75
32B34.7 GB33 tok/s33K ctx
dense
👁 Tsinghua/Zhipu
CodeGeeX 4 9B
A74
9B17.4 GB75 tok/s131K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 70B
A74
70B58.9 GB15 tok/s50K ctx
dense
👁 BAAI
BGE M3
A74
0.57B13.5 GB8 tok/s8K ctx
dense
👁 Meta
CodeLlama 13B Instruct
A74
13B31.4 GB38 tok/s16K ctx
dense
👁 MosaicML
MPT-30B-Instruct
A74
30B56.3 GB28 tok/s8K ctx
dense
👁 Google
Gemma 4 E4B
A73
8B17.4 GB63 tok/s128K ctx
dense
👁 AllenAI
OLMo 2 13B
A73
13B21.6 GB41 tok/s33K ctx
dense
👁 Alibaba
Qwen 2.5 7B
A73
7B16.4 GB95 tok/s131K ctx
dense
👁 DevStral AI
DevStral 7B
A72
7B17.5 GB94 tok/s8K ctx
dense
👁 Meta
CodeLlama 7B Instruct
A72
7B23.4 GB88 tok/s16K ctx
dense
👁 IBM
Granite Code 8B
A72
8B18.1 GB83 tok/s8K ctx
dense
👁 BigCode
StarCoder 7B
A72
7B22.9 GB88 tok/s8K ctx
dense
👁 LLaVA
LLaVA 1.6 13B
A72
13B31.4 GB38 tok/s4K ctx
dense
👁 Mistral AI
Codestral Mamba 7B
A72
7B16.0 GB93 tok/s262K ctx
state-space
👁 IBM
Granite 4.1 8B
A72
8B18.6 GB83 tok/s131K ctx
dense
👁 TII
Falcon 40B Instruct
A71
40B41.9 GB23 tok/s8K ctx
dense
👁 Mixedbread AI
mxbai Embed Large
A71
0.34B13.4 GB5 tok/s1K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 14B
A71
14B22.7 GB38 tok/s33K ctx
dense
👁 Nous Research
Nous Hermes 1.0
A71
9B29.0 GB68 tok/s16K ctx
dense
👁 Cohere
Command A 111B
A70
111B82.9 GB7 tok/s4K ctx
dense
👁 NVIDIA
Nemotron 70B
A70
70B58.9 GB17 tok/s50K ctx
dense
👁 Snowflake
Snowflake Arctic Embed L
A70
0.34B13.4 GB5 tok/s1K ctx
dense
B

B Tier -- Good Fit (54 models)

Solid performance, may need quantization or reduced context.

👁 Alibaba
Qwen 2.5 Coder 3B
B70
3B15.3 GB42 tok/s131K ctx
dense
👁 Mistral AI
Pixtral 12B
B70
12B21.0 GB45 tok/s131K ctx
dense
👁 InternLM
InternLM 7B
B70
7B23.4 GB88 tok/s8K ctx
dense
👁 InternLM
InternLM Chat 7B
B70
7B23.4 GB88 tok/s8K ctx
dense
👁 Google
Gemma 2 27B
B70
27B39.0 GB27 tok/s8K ctx
dense
👁 WizardLM
WizardLM 13B
B69
13B31.4 GB38 tok/s8K ctx
dense
👁 LMSYS
Vicuna 13B
B69
13B31.4 GB38 tok/s4K ctx
dense
👁 Nous Research
Nous Dolphin 13B
B69
13B32.8 GB33 tok/s16K ctx
dense
👁 Google
Gemma 4 E2B
B69
5.1B14.9 GB71 tok/s128K ctx
dense
B68
7B18.5 GB94 tok/s8K ctx
dense
👁 BAAI
BGE Large EN v1.5
B68
0.34B13.4 GB5 tok/s1K ctx
dense
👁 Meta
Llama 3.1 8B
B68
8B18.1 GB83 tok/s128K ctx
dense
👁 Zhipu
GLM-4 9B
B68
9B17.4 GB75 tok/s128K ctx
dense
👁 WizardLM
WizardMath 7B
B68
7B17.5 GB94 tok/s4K ctx
dense
👁 Allen AI
OLMo 2 7B
B68
7B17.5 GB94 tok/s4K ctx
dense
👁 LLaVA
LLaVA 1.5 7B
B67
7B23.4 GB88 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 7B
B67
7B16.4 GB95 tok/s131K ctx
dense
👁 Nomic AI
Nomic Embed Text v1.5
B67
0.14B12.1 GB2 tok/s8K ctx
dense
👁 Google
Gemma 3 4B
B66
4B15.8 GB56 tok/s128K ctx
dense
👁 Mistral
Ministral 3 3B
B66
3B13.8 GB42 tok/s262K ctx
multimodal
👁 Mistral
Mixtral 8x7B
B66
47B41.9 GB25 tok/s33K ctx
moe
👁 Microsoft
Phi 4 Mini 4B
B66
4B15.2 GB56 tok/s128K ctx
dense
👁 MosaicML
MPT-7B-Instruct
B65
7B23.4 GB88 tok/s8K ctx
dense
👁 Baichuan
Baichuan 7B
B65
7B23.4 GB88 tok/s8K ctx
dense
👁 Baichuan
Baichuan 13B
B65
13B32.8 GB33 tok/s8K ctx
dense
👁 Meta
Llama 4 Scout 17B 16E
B65
109B80.7 GB10 tok/s4K ctx
moe
👁 Alibaba
Qwen 3.5 2B
B65
2B14.2 GB28 tok/s131K ctx
dense
👁 Cerebras
Cerebras-GPT 13B
B64
13B30.4 GB33 tok/s79K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 7B
B64
7B16.4 GB95 tok/s33K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 8B
B64
8B18.1 GB83 tok/s33K ctx
dense
7B17.5 GB94 tok/s4K ctx
dense
👁 Microsoft
Phi 3 Mini 3.8B
B63
3.8B19.4 GB53 tok/s128K ctx
dense
👁 Alibaba
Qwen 2.5 3B
B63
3B15.3 GB42 tok/s131K ctx
dense
👁 TII
Falcon 7B Instruct
B63
7B15.7 GB97 tok/s8K ctx
dense
👁 01.AI
Yi 1.5 34B
B62
34B35.7 GB31 tok/s4K ctx
dense
👁 Microsoft
Phi 3.5 Mini 4B
B62
4B19.6 GB56 tok/s128K ctx
dense
👁 Alibaba
Qwen 3 1.7B
B62
1.7B14.0 GB24 tok/s33K ctx
dense
👁 Google
Gemma 2 9B
B61
9B21.9 GB54 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 Math 72B
B61
72B60.1 GB15 tok/s4K ctx
dense
B61
0.02B11.6 GB2 tok/s0K ctx
dense
👁 IBM
Granite 4.1 3B
B61
3B14.3 GB42 tok/s131K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 14B
B60
14B22.7 GB38 tok/s131K ctx
dense
👁 Meta
Llama 3.2 11B Vision
B60
11B19.9 GB49 tok/s16K ctx
vision
👁 IBM
Granite Code 3B
B60
3B15.6 GB42 tok/s8K ctx
dense
👁 InternLM
InternLM 20B
B60
20B44.0 GB31 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 1.5B
B60
1.5B12.6 GB21 tok/s33K ctx
dense
👁 Mistral
Mistral 7B Instruct v0.3
B60
7B17.5 GB94 tok/s8K ctx
dense
👁 01.AI
Yi Coder 9B
B59
9B18.2 GB74 tok/s131K ctx
dense
👁 Mistral
Mistral Nemo 12B
B58
12B21.0 GB45 tok/s128K ctx
dense
👁 DeepSeek
DeepSeek LLM 67B
B58
67B57.9 GB16 tok/s4K ctx
dense
👁 Mistral AI
Codestral 22B
B57
22B27.1 GB37 tok/s33K ctx
dense
👁 Microsoft
Phi 3 Medium 14B
B57
14B22.9 GB38 tok/s128K ctx
dense
👁 Mistral
Ministral 8B
B57
8B18.3 GB83 tok/s131K ctx
dense
👁 Meta
Llama 3.2 3B
B56
3B14.8 GB42 tok/s128K ctx
dense
C

C Tier -- Tight Fit (193 models)

Runs with offloading or significant compromises.

👁 Cohere
Aya Expanse 32B
C54
32B33.2 GB34 tok/s8K ctx
dense
👁 Cohere
Command R+ 104B
C54
104B78.1 GB9 tok/s4K ctx
dense
👁 IBM
Granite 3.1 8B
C52
8B18.1 GB87 tok/s128K ctx
state-space
C52
3B15.1 GB42 tok/s128K ctx
dense
👁 DeepSeek
DeepSeek R1 1.5B
C52
1.5B12.6 GB21 tok/s33K ctx
dense
👁 TinyLlama
TinyLlama 1.1B
C52
1.1B12.3 GB15 tok/s4K ctx
dense
👁 01.AI
Yi 1.5 9B
C51
9B18.2 GB74 tok/s4K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
C51
48B46.2 GB21 tok/s81K ctx
dense
👁 Alibaba
Qwen 2.5 Math 7B
C51
7B16.4 GB95 tok/s4K ctx
dense
👁 01.AI
Yi 34B Chat
C51
34B35.7 GB31 tok/s162K ctx
dense
👁 Unsloth
Qwen3.5 35B A3B
C50
35B36.7 GB28 tok/s142K ctx
dense
Qwen3.5 35B A3B
C50
35B36.7 GB28 tok/s142K ctx
dense
👁 LMSYS
Vicuna 7B
C50
7B23.4 GB88 tok/s4K ctx
dense
Hermes 4.3 36B
C49
36B37.4 GB27 tok/s136K ctx
dense
👁 OpenChat
OpenChat 7B
C49
7B17.5 GB94 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 1.5B
C49
1.5B12.6 GB21 tok/s131K ctx
dense
EXAONE 4.0 32B
C49
32B34.5 GB31 tok/s164K ctx
dense
aya expanse 32b heretic MPOA i1
C49
32B34.5 GB31 tok/s164K ctx
dense
👁 Cohere
Aya Expanse 8B
C49
8B18.1 GB83 tok/s8K ctx
dense
Baichuan M2 32B Q4 K M
C49
32B34.5 GB31 tok/s164K ctx
dense
baichuan inc Baichuan M2 32B
C49
32B34.5 GB31 tok/s164K ctx
dense
👁 Unsloth
Qwen3.5 27B
C49
27B30.9 GB33 tok/s209K ctx
dense
👁 Unsloth
gemma 3 27b it
C48
27B30.9 GB33 tok/s209K ctx
dense
gemma 3 27b it
C48
27B30.9 GB33 tok/s209K ctx
dense
👁 Stability AI
StableLM 2 12B
C48
12B32.1 GB33 tok/s4K ctx
dense
👁 Teknium
OpenHermes 2.5 7B
C48
7B17.5 GB94 tok/s8K ctx
dense
Llama 3.3 70B Instruct
C48
70B62.2 GB14 tok/s30K ctx
dense
👁 Hugging Face H4
Zephyr 7B Beta
C48
7B17.5 GB94 tok/s33K ctx
dense
8B18.1 GB83 tok/s33K ctx
dense
👁 Unsloth
Mistral Small 3.2 24B Instruct 2506
C48
24B28.7 GB34 tok/s246K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C48
24B28.7 GB34 tok/s246K ctx
dense
Solar Open 69B REAP i1
C48
69B61.4 GB14 tok/s31K ctx
dense
Mistral Small 24B Instruct 2501
C48
24B28.7 GB34 tok/s246K ctx
dense
👁 Google
Gemma 3 1B
C48
1B12.3 GB14 tok/s33K ctx
dense
👁 UC Berkeley
Starling LM 7B
C48
7B17.5 GB94 tok/s8K ctx
dense
mistral small 3.1 24b instruct 2503 hf
C48
24B28.7 GB34 tok/s246K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
C48
24B28.7 GB34 tok/s246K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
C47
24B28.7 GB34 tok/s246K ctx
dense
stabilityai japanese stablelm instruct beta 70b
C47
70B62.2 GB14 tok/s30K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C47
24B28.7 GB34 tok/s246K ctx
dense
BaichuanMed OCR 72B i1
C47
72B63.6 GB14 tok/s26K ctx
dense
👁 DeepSeek
DeepSeek LLM 7B
C47
7B22.9 GB88 tok/s4K ctx
dense
Codestral 22B v0.1
C47
22B27.3 GB35 tok/s276K ctx
dense
Llama 2 7B Chat
C47
7B16.4 GB88 tok/s1045K ctx
dense
👁 BigCode
StarCoder2 15B
C47
15B23.3 GB33 tok/s16K ctx
dense
Codestral 22B v0.1
C47
22B27.3 GB35 tok/s276K ctx
dense
👁 Google
Gemma 2 2B
C47
2B14.1 GB28 tok/s8K ctx
dense
Mistral 7B Instruct v0.2
C47
7B16.4 GB88 tok/s1045K ctx
dense
Mistral 7B Instruct v0.3
C47
7B16.4 GB88 tok/s1045K ctx
dense
Codestral 22B v0.1
C47
22B27.3 GB35 tok/s276K ctx
dense
Meta Llama 3.1 8B Instruct
C47
8B17.1 GB77 tok/s904K ctx
dense
llava llama 3 8b v1 1
C47
8B17.1 GB77 tok/s904K ctx
dense
👁 Unsloth
DeepSeek R1 0528 Qwen3 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
Mistral 7B Instruct v0.3
C47
7B16.4 GB88 tok/s1045K ctx
dense
👁 Unsloth
Qwen3.5 9B
C47
9B17.8 GB68 tok/s794K ctx
dense
Hermes 2 Pro Mistral 7B
C47
7B16.4 GB88 tok/s1045K ctx
dense
Meta Llama 3 8B Instruct
C47
8B17.1 GB77 tok/s904K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Llama 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
Codestral 22B v0.1 IMat
C47
22B27.3 GB35 tok/s276K ctx
dense
falcon mamba 7b instruct Q4 K M
C47
7B16.4 GB93 tok/s1045K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
C47
9B17.8 GB68 tok/s794K ctx
dense
Dolphin3.0 Llama3.1 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
Llama 3 8B Instruct 32k v0.1
C47
8B17.1 GB77 tok/s904K ctx
dense
zephyr 7B beta
C47
7B16.4 GB88 tok/s1045K ctx
dense
Meta Llama 3.1 8B Instruct
C47
8B17.1 GB77 tok/s904K ctx
dense
vntl llama3 8b v2
C47
8B17.1 GB77 tok/s904K ctx
dense
Nous Hermes 2 Mistral 7B DPO
C47
7B16.4 GB88 tok/s1045K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
Falcon H1R 7B
C47
7B16.4 GB88 tok/s1045K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
Qwen3.5 9B
C47
9B17.8 GB68 tok/s794K ctx
dense
zephyr 7B alpha
C47
7B16.4 GB88 tok/s1045K ctx
dense
speechless zephyr code functionary 7b
C47
7B16.4 GB88 tok/s1045K ctx
dense
Codestral 22B v0.1 i1
C47
22B27.3 GB35 tok/s276K ctx
dense
Qwen3 8B DeepSeek v3.2 Speciale Distill
C47
8B17.1 GB77 tok/s904K ctx
dense
StarCoder2 7B
C47
7B16.4 GB88 tok/s1045K ctx
dense
Hermes 3 Llama 3.1 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
japanese stablelm instruct gamma 7B
C47
7B16.4 GB88 tok/s1045K ctx
dense
Mamba Codestral 7B v0.1
C47
7B16.4 GB93 tok/s1045K ctx
dense
Falcon H1 7B Instruct
C47
7B16.4 GB88 tok/s1045K ctx
dense
👁 Unsloth
Falcon H1R 7B
C47
7B16.4 GB88 tok/s1045K ctx
dense
Hermes 2 Pro Llama 3 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
👁 Intel
Neural Chat 7B
C47
7B17.5 GB94 tok/s8K ctx
dense
Dolphin3.0 Llama3.1 8B
C47
8B17.1 GB77 tok/s904K ctx
dense
granite 8b code instruct 4k
C47
8B17.1 GB77 tok/s904K ctx
dense
Yi 1.5 6B Chat
C47
6B15.6 GB84 tok/s1233K ctx
dense
Codestral 21B Pruned i1
C47
21B26.5 GB35 tok/s293K ctx
dense
dolphin 2.9.4 llama3.1 8b
C47
8B17.1 GB77 tok/s904K ctx
dense
Yi Coder 9B Chat
C47
9B17.8 GB68 tok/s794K ctx
dense
glm 4 9b chat 1m
C46
9B17.8 GB68 tok/s794K ctx
dense
stabilityai japanese stablelm base gamma 7b
C46
7B16.4 GB88 tok/s1045K ctx
dense
baichuan2 7b chat
C46
7B16.4 GB88 tok/s1045K ctx
dense
EXAONE 3.5 7.8B Instruct
C46
7.8B16.9 GB79 tok/s929K ctx
dense
stablelm 2 zephyr 1 6b
C46
6B15.6 GB84 tok/s1233K ctx
dense
openchat 3.6 8b 20240522 IMat
C46
8B17.1 GB77 tok/s904K ctx
dense
DiscoPOP zephyr 7b gemma
C46
7B16.4 GB88 tok/s1045K ctx
dense
internlm2 5 20b chat
C46
20B25.8 GB36 tok/s312K ctx
dense
aya expanse 8b
C46
8B17.1 GB77 tok/s904K ctx
dense
EXAONE 3.5 7.8B Instruct
C46
7.8B16.9 GB79 tok/s929K ctx
dense
starcoder2 7b
C46
7B16.4 GB88 tok/s1045K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
C46
7B16.4 GB88 tok/s1045K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
internlm2 limarp chat 20b
C46
20B25.8 GB36 tok/s312K ctx
dense
EXAONE 3.5 7.8B Instruct i1
C46
7.8B16.9 GB79 tok/s929K ctx
dense
internlm2 math plus 7b IMat
C46
7B16.4 GB88 tok/s1045K ctx
dense
dolphin v2 8b abliterated i1
C46
8B17.1 GB77 tok/s904K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 0.5B
C46
0.5B11.8 GB7 tok/s131K ctx
dense
exaone 3.0 7.8b it
C46
7.8B16.9 GB79 tok/s929K ctx
dense
internlm2 5 7b chat i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
Yi 1.5 9B Chat
C46
9B17.8 GB68 tok/s794K ctx
dense
👁 01.AI
Yi 1.5 6B
C46
6B15.9 GB84 tok/s4K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
C46
7B16.4 GB88 tok/s1045K ctx
dense
MD Judge v0 2 internlm2 7b i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
internlm2 math plus 20b i1
C46
20B25.8 GB36 tok/s312K ctx
dense
aya expanse 8b orthogonal heretic i1
C46
8B17.1 GB77 tok/s904K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
C46
7B16.4 GB88 tok/s1045K ctx
dense
jointpreferences mistral 7b sft helpful
C46
7B16.4 GB88 tok/s1045K ctx
dense
zephyr 7b dpo full i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
blossom v3 baichuan2 7b i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
blossom v1 baichuan 7b i1
C46
7B16.4 GB88 tok/s1045K ctx
dense
Yi 1.5 6B Chat
C46
6B15.6 GB84 tok/s1233K ctx
dense
stablelm 2 1 6b chat imatrix
C46
6B15.6 GB84 tok/s1233K ctx
dense
aya expanse 8b orthogonal heretic
C46
8B17.1 GB77 tok/s904K ctx
dense
internlm2 5 1 8b chat i1
C46
8B17.1 GB77 tok/s904K ctx
dense
internlm3 8b instruct abliterated i1
C46
8B17.1 GB77 tok/s904K ctx
dense
Yi 9B Coder i1
C46
9B17.8 GB68 tok/s794K ctx
dense
Codestral RAG 19B Pruned i1
C46
19B25.1 GB36 tok/s332K ctx
dense
👁 BigCode
StarCoder2 7B
C46
7B16.0 GB96 tok/s16K ctx
dense
HelpingAI2 9B
C46
9B17.8 GB68 tok/s794K ctx
dense
HelpingAI2 6B
C46
6B15.6 GB84 tok/s1233K ctx
dense
HelpingAI2 9B i1
C46
9B17.8 GB68 tok/s794K ctx
dense
HelpingAI 9B 200k i1
C46
9B17.8 GB68 tok/s794K ctx
dense
HelpingAI2 6B i1
C46
6B15.6 GB84 tok/s1233K ctx
dense
👁 NVIDIA
Nemotron Mini 4B
C46
4B15.7 GB56 tok/s4K ctx
dense
HelpingAI 9B i1
C46
9B17.8 GB68 tok/s794K ctx
dense
HelpingAI2.5 10B i1
C46
10B18.5 GB62 tok/s707K ctx
dense
gemma 3 12b it
C46
12B20.0 GB41 tok/s575K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Qwen 14B
C46
14B21.4 GB35 tok/s481K ctx
dense
Phi 4 reasoning vision 15B
C46
15B22.2 GB35 tok/s443K ctx
dense
StarCoder2 15B
C46
15B22.2 GB35 tok/s443K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C46
10.7B19.0 GB46 tok/s655K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C45
10.7B19.0 GB46 tok/s655K ctx
dense
NousResearch Hermes 4 14B
C45
14B21.4 GB35 tok/s481K ctx
dense
starcoder2 15b instruct v0.1
C45
15B22.2 GB35 tok/s443K ctx
dense
starcoder2 15b i1
C45
15B22.2 GB35 tok/s443K ctx
dense
starcoder2 15b instruct v0.1
C45
15B22.2 GB35 tok/s443K ctx
dense
solar finalised finetuned Model 10.7B i1
C45
10.7B19.0 GB46 tok/s655K ctx
dense
SOLAR 10.7B v1.0
C45
10.7B19.0 GB46 tok/s655K ctx
dense
HelpingAI 15B i1
C45
15B22.2 GB35 tok/s443K ctx
dense
HelpingAI2.5 5B i1
C45
5B14.9 GB70 tok/s1496K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
C45
14B21.4 GB35 tok/s481K ctx
dense
Helply 10.2b chat i1
C45
10.2B18.7 GB49 tok/s691K ctx
dense
internlm JanusCoder 14B
C45
14B21.4 GB35 tok/s481K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
C45
14B21.4 GB35 tok/s481K ctx
dense
👁 Unsloth
Qwen3.5 4B
C45
4B14.2 GB56 tok/s1891K ctx
dense
gemma 3 4b it
C45
4B14.2 GB56 tok/s1891K ctx
dense
gemma 3 4b it
C45
4B14.2 GB56 tok/s1891K ctx
dense
👁 Alibaba
Qwen 3 0.6B
C44
0.6B12.5 GB8 tok/s33K ctx
dense
👁 Meta
Llama 3.2 1B
C44
1B12.4 GB14 tok/s128K ctx
dense
Llama 3.2 3B Instruct
C44
3B13.8 GB42 tok/s2535K ctx
dense
👁 Qwen
Qwen2.5 3B Instruct
C44
3B13.4 GB42 tok/s2550K ctx
dense
Llama 3.2 3B Instruct
C44
3B13.4 GB42 tok/s2550K ctx
dense
Ministral 3 3B Instruct 2512
C44
3B13.4 GB42 tok/s2550K ctx
dense
HELVETE 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
Hermes 3 Llama 3.2 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
stablelm zephyr 3b
C43
3B13.4 GB42 tok/s2550K ctx
dense
StarCoder2 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
AI21 Jamba Reasoning 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
stablelm 3b 4e1t
C43
3B13.4 GB42 tok/s2550K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
ai21labs AI21 Jamba2 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
HelpingAI 3B hindi i1
C43
3B13.4 GB42 tok/s2550K ctx
dense
AI21 Jamba2 3B
C43
3B13.4 GB42 tok/s2550K ctx
dense
HelpingAI 3B hindi
C43
3B13.4 GB42 tok/s2550K ctx
dense
AI21 Jamba2 3B i1
C43
3B13.4 GB42 tok/s2550K ctx
dense
EXAONE 3.5 2.4B Instruct
C42
2.4B13.0 GB34 tok/s3208K ctx
dense
gemma 2 2b it
C42
2B13.1 GB28 tok/s3837K ctx
dense
👁 Google
gemma 2b
C42
2B12.7 GB28 tok/s3866K ctx
dense
Gemmasutra Mini 2B v1
C42
2B12.7 GB28 tok/s3866K ctx
dense
👁 BigCode
StarCoder2 3B
C42
3B13.6 GB42 tok/s16K ctx
dense
gemma 2 2b it
C42
2B12.7 GB28 tok/s3866K ctx
dense
👁 Alibaba
Qwen 3.5 0.6B
C42
0.6B12.5 GB8 tok/s131K ctx
dense
👁 Qwen
Qwen2.5 1.5B Instruct
C41
1.5B12.4 GB21 tok/s5183K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Qwen 1.5B
C41
1.5B12.4 GB21 tok/s5183K ctx
dense
Yi Coder 1.5B Chat
C41
1.5B12.4 GB21 tok/s5183K ctx
dense
stablelm 2 zephyr 1.6b
C41
1.6B12.4 GB22 tok/s4853K ctx
dense
logos16v2 stablelm2 1.6b i1
C41
1.6B12.4 GB22 tok/s4853K ctx
dense
Yi Coder 1.5B
C41
1.5B12.4 GB21 tok/s5183K ctx
dense
👁 Unsloth
Falcon H1 1.5B Instruct
C41
1.5B12.4 GB21 tok/s5183K ctx
dense
TinyLlama 1.1B Chat v1.0
C40
1.1B12.1 GB15 tok/s7097K ctx
dense
👁 Alibaba
Qwen 2.5 0.5B
C40
0.5B11.8 GB7 tok/s131K ctx
dense
F

F Tier -- Can't Run (45 models)

Too heavy for this hardware.

EXAONE 4.0 1.2B
D40
1.2B12.1 GB17 tok/s6499K ctx
dense
Llama 3.2 1B Instruct Q8 0
D40
1B12.2 GB14 tok/s7787K ctx
dense
TinyLlama 1.1B Chat v0.3
D40
1.1B12.1 GB15 tok/s7097K ctx
dense
👁 TinyLlama
TinyLlama 1.1B Chat v0.6
D40
1.1B12.1 GB15 tok/s7097K ctx
dense
Llama 3.2 1B Instruct
D40
1B12.0 GB14 tok/s7815K ctx
dense
gemma 3 1b it
D40
1B12.0 GB14 tok/s7815K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
D39
1.1B12.1 GB15 tok/s7097K ctx
dense
Falcon3 1B Instruct abliterated
D39
1B12.0 GB14 tok/s7815K ctx
dense
SmolVLM 500M Instruct
D38
0.5B11.8 GB7 tok/s9191K ctx
dense
embeddinggemma 300M
D36
0.3B11.6 GB4 tok/s9217K ctx
dense
granite embedding 107m multilingual
D34
0.11B11.4 GB2 tok/s9246K ctx
dense
Falcon H1 Tiny 90M Instruct
D34
0.09B11.4 GB2 tok/s9248K ctx
dense
👁 Alibaba
Qwen 3 235B A22B
F0
235B157.5 GB3 tok/s4K ctx
moe
👁 Alibaba
Qwen3-Coder 480B A35B Instruct
F0
480B306.9 GB3 tok/s4K ctx
moe
👁 Alibaba
Qwen 3.5 122B A10B
F0
122B88.1 GB15 tok/s4K ctx
moe
👁 Alibaba
Qwen 3.5 397B A17B
F0
397B256.3 GB3 tok/s4K ctx
moe
👁 Mistral
Mixtral 8x22B
F0
141B100.7 GB5 tok/s4K ctx
moe
👁 Mistral
Devstral 2 123B Instruct
F0
123B91.7 GB6 tok/s4K ctx
dense
👁 Mistral
Mistral Small 4 119B
F0
119B89.2 GB16 tok/s4K ctx
moe
👁 Mistral
Leanstral 119B A6B
F0
119B92.6 GB14 tok/s4K ctx
moe
👁 Mistral
Mistral Large 3
F0
675B428.4 GB2 tok/s4K ctx
+1moe
F0
744B484.2 GB2 tok/s4K ctx
moe
👁 Z.ai
GLM-5.1
F0
754B490.3 GB2 tok/s4K ctx
moe
F0
1000B628.7 GB2 tok/s4K ctx
moe
F0
1000B628.7 GB2 tok/s4K ctx
+1moe
👁 DeepSeek
DeepSeek V3 671B
F0
671B480.1 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V3.1 671B
F0
671B480.1 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V4 Pro
F0
1600B875.1 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V4 Flash
F0
284B170.6 GB5 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek R1 671B
F0
671B480.1 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek Coder V2 236B
F0
236B213.8 GB5 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V2.5 236B
F0
236B213.8 GB5 tok/s4K ctx
moe
👁 Meta
Llama 4 Maverick 17B 128E
F0
400B258.2 GB3 tok/s4K ctx
moe
👁 Mistral AI
Pixtral Large 124B
F0
124B92.3 GB6 tok/s4K ctx
dense
👁 Meta
Llama 3.1 405B
F0
405B266.0 GB2 tok/s4K ctx
dense
👁 OpenAI
GPT-OSS 120B
F0
117B87.5 GB7 tok/s4K ctx
dense
👁 DeepSeek
DeepSeek V3.2
F0
671B421.0 GB2 tok/s4K ctx
moe
MiniMax M2.7
F0
230B155.4 GB6 tok/s4K ctx
moe
Solar Open 100B
F0
100B84.0 GB7 tok/s4K ctx
dense
K EXAONE 236B A23B
F0
236B182.9 GB2 tok/s4K ctx
dense
Baichuan M3 235B
F0
235B182.2 GB2 tok/s4K ctx
dense
Baichuan M3 235B i1
F0
235B182.2 GB2 tok/s4K ctx
dense
Solar Open 100B i1
F0
100B84.0 GB7 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 122B A10B
F0
122B85.3 GB7 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 397B A17B
F0
397B300.0 GB2 tok/s4K ctx
dense

Compare with...

See How Other Hardware Stacks Up

Check tier lists for similar GPUs and Apple Silicon configs

Mac Studio M3 Ultra 96GB

96 GB unified memory

View tier list

MacBook Pro M2 Max 96GB

96 GB unified memory

View tier list

MacBook Pro M1 Max 64GB

64 GB unified memory

View tier list

Mac Studio M2 Ultra 64GB

64 GB unified memory

View tier list
Full MacBook Pro M4 Max 96GB pageCompare this hardware