NVIDIA GH200 96GB

72B62.9 GB74 tok/s79K ctx

C Tier -- Tight Fit (186 models)

Runs with offloading or significant compromises.

👁 Meta
Llama 3.2 3B

C55

3B14.0 GB42 tok/s128K ctx

dense

BaichuanMed OCR 72B i1

C55

69B60.7 GB77 tok/s86K ctx

Solar Open 69B REAP i1

C55

stabilityai japanese stablelm instruct beta 70b

C55

70B61.4 GB76 tok/s83K ctx

100B83.2 GB53 tok/s33K ctx

👁 Cohere
Aya Expanse 32B

C54

32B32.5 GB181 tok/s8K ctx

dense

👁 Unsloth
Qwen3.5 122B A10B

C52

122B84.6 GB50 tok/s29K ctx

dense

Solar Open 100B

C52

100B83.2 GB53 tok/s33K ctx

Solar Open 100B i1

C52

Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored

C52

48B45.4 GB111 tok/s160K ctx

35B36.0 GB152 tok/s250K ctx

👁 01.AI
Yi 34B Chat

C51

34B34.9 GB170 tok/s200K ctx

dense

👁 DeepSeek
DeepSeek R1 1.5B

C50

1.5B11.8 GB21 tok/s33K ctx

dense

👁 TinyLlama
TinyLlama 1.1B

C50

1.1B11.5 GB15 tok/s4K ctx

dense

👁 IBM
Granite 3.1 8B

C50

8B17.3 GB112 tok/s128K ctx

state-space

👁 HuggingFace
SmolLM3 3B

C50

3B14.3 GB42 tok/s128K ctx

dense

👁 Unsloth
Qwen3.5 35B A3B

C50

35B36.0 GB152 tok/s250K ctx

dense

👁 01.AI
Yi 1.5 9B

C50

9B17.5 GB126 tok/s4K ctx

dense

Qwen3.5 35B A3B

C50

36B36.7 GB148 tok/s241K ctx

Hermes 4.3 36B

C50

👁 Alibaba
Qwen 2.5 Math 7B

C50

7B15.6 GB98 tok/s4K ctx

dense

EXAONE 4.0 32B

aya expanse 32b heretic MPOA i1

Baichuan M2 32B Q4 K M

baichuan inc Baichuan M2 32B

👁 Unsloth
Qwen3.5 27B

C49

27B30.1 GB197 tok/s349K ctx

dense

👁 Unsloth
gemma 3 27b it

C49

27B30.1 GB197 tok/s349K ctx

dense

gemma 3 27b it

27B30.1 GB197 tok/s349K ctx

cognitivecomputations Dolphin Mistral 24B Venice Edition

👁 Stability AI
StableLM 2 12B

C49

12B31.3 GB168 tok/s4K ctx

dense

👁 BigCode
StarCoder2 15B

C48

15B22.5 GB210 tok/s16K ctx

dense

👁 Unsloth
Mistral Small 3.2 24B Instruct 2506

C48

24B28.0 GB221 tok/s403K ctx

dense

Mistral Small 24B Instruct 2501

mistral small 3.1 24b instruct 2503 hf

cognitivecomputations Dolphin3.0 R1 Mistral 24B

Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1

cognitivecomputations Dolphin Mistral 24B Venice Edition

👁 Alibaba
Qwen 2.5 1.5B

C48

1.5B11.8 GB21 tok/s131K ctx

dense

Codestral 22B v0.1

Codestral 22B v0.1

👁 OpenChat
OpenChat 7B

C48

7B16.7 GB98 tok/s8K ctx

dense

Codestral 22B v0.1

👁 LMSYS
Vicuna 7B

C47

7B22.6 GB98 tok/s4K ctx

dense

Codestral 22B v0.1 IMat

👁 Cohere
Aya Expanse 8B

C47

8B17.3 GB112 tok/s8K ctx

dense

Codestral 22B v0.1 i1

Codestral 21B Pruned i1

21B25.8 GB253 tok/s473K ctx

internlm2 5 20b chat

20B25.0 GB266 tok/s500K ctx

internlm2 limarp chat 20b

20B25.0 GB266 tok/s500K ctx

internlm2 math plus 20b i1

20B25.0 GB266 tok/s500K ctx

Codestral RAG 19B Pruned i1

19B24.3 GB266 tok/s531K ctx

Phi 4 reasoning vision 15B

👁 Unsloth
DeepSeek R1 Distill Qwen 14B

C47

14B20.7 GB196 tok/s751K ctx

dense

StarCoder2 15B

NousResearch Hermes 4 14B

👁 Google
Gemma 3 1B

C46

1B11.5 GB14 tok/s33K ctx

dense

👁 Teknium
OpenHermes 2.5 7B

C46

7B16.7 GB98 tok/s8K ctx

dense

👁 Cognitive Computations
Dolphin 2.9 8B

C46

8B17.3 GB112 tok/s33K ctx

dense

starcoder2 15b instruct v0.1

starcoder2 15b i1

starcoder2 15b instruct v0.1

gemma 3 12b it

12B19.2 GB168 tok/s890K ctx

HelpingAI 15B i1

GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV

👁 Hugging Face H4
Zephyr 7B Beta

C46

7B16.7 GB98 tok/s33K ctx

dense

internlm JanusCoder 14B

GGUF SOLARized GraniStral 14B 1902 YeAM HCT

Qwen3.5 9B Uncensored HauhauCS Aggressive

👁 Unsloth
Qwen3.5 9B

C46

9B17.0 GB126 tok/s1214K ctx

dense

SOLAR 10.7B Instruct v1.0 uncensored

Qwen3.5 9B

Nous Hermes 2 SOLAR 10.7B

👁 UC Berkeley
Starling LM 7B

C46

7B16.7 GB98 tok/s8K ctx

dense

Meta Llama 3.1 8B Instruct

llava llama 3 8b v1 1

👁 Unsloth
DeepSeek R1 0528 Qwen3 8B

C46

8B16.3 GB112 tok/s1376K ctx

dense

Yi Coder 9B Chat

Meta Llama 3 8B Instruct

👁 Google
Gemma 2 2B

C46

2B13.3 GB28 tok/s8K ctx

dense

👁 Unsloth
DeepSeek R1 Distill Llama 8B

C46

8B16.3 GB112 tok/s1376K ctx

dense

glm 4 9b chat 1m

Dolphin3.0 Llama3.1 8B

Llama 3 8B Instruct 32k v0.1

Meta Llama 3.1 8B Instruct

Llama 2 7B Chat

vntl llama3 8b v2

DeepSeek R1 0528 Qwen3 8B

Mistral 7B Instruct v0.2

solar finalised finetuned Model 10.7B i1

DeepSeek R1 0528 Qwen3 8B

Mistral 7B Instruct v0.3

Yi 1.5 9B Chat

SOLAR 10.7B v1.0

Qwen3 8B DeepSeek v3.2 Speciale Distill

Hermes 3 Llama 3.1 8B

Hermes 2 Pro Llama 3 8B

Dolphin3.0 Llama3.1 8B

granite 8b code instruct 4k

dolphin 2.9.4 llama3.1 8b

Mistral 7B Instruct v0.3

Hermes 2 Pro Mistral 7B

openchat 3.6 8b 20240522 IMat

HelpingAI2.5 10B i1

10B17.8 GB140 tok/s1084K ctx

Helply 10.2b chat i1

10.2B17.9 GB143 tok/s1061K ctx

Yi 9B Coder i1

zephyr 7B beta

aya expanse 8b

EXAONE 3.5 7.8B Instruct

Nous Hermes 2 Mistral 7B DPO

dolphin v2 8b abliterated i1

EXAONE 3.5 7.8B Instruct

HelpingAI2 9B

Falcon H1R 7B

EXAONE 3.5 7.8B Instruct i1

aya expanse 8b orthogonal heretic i1

falcon mamba 7b instruct Q4 K M

exaone 3.0 7.8b it

zephyr 7B alpha

speechless zephyr code functionary 7b

HelpingAI2 9B i1

HelpingAI 9B 200k i1

👁 DeepSeek
DeepSeek LLM 7B

C45

7B22.1 GB98 tok/s4K ctx

dense

👁 Alibaba
Qwen 2.5 Coder 0.5B

C45

0.5B11.0 GB7 tok/s131K ctx

dense

StarCoder2 7B

HelpingAI 9B i1

japanese stablelm instruct gamma 7B

aya expanse 8b orthogonal heretic

Falcon H1 7B Instruct

👁 Unsloth
Falcon H1R 7B

C45

7B15.6 GB98 tok/s1584K ctx

dense

internlm2 5 1 8b chat i1

internlm3 8b instruct abliterated i1

Yi 1.5 6B Chat

Mamba Codestral 7B v0.1

stabilityai japanese stablelm base gamma 7b

baichuan2 7b chat

DiscoPOP zephyr 7b gemma

stablelm 2 zephyr 1 6b

starcoder2 7b

zephyr 7b beta Mistral 7B Instruct v0.2

OpenChat 3.5 7B Qwen v2.0 i1

OpenChat 3.5 7B Starling v2.0 i1

internlm2 math plus 7b IMat

CodeNinja 1.0 OpenChat 7B i1

👁 Intel
Neural Chat 7B

C45

7B16.7 GB98 tok/s8K ctx

dense

internlm2 5 7b chat i1

OpenSafetyLab MD Judge v0 2 internlm2 7b

MD Judge v0 2 internlm2 7b i1

zephyr 7b gemma sft african ultrachat 100k

jointpreferences mistral 7b sft helpful

zephyr 7b dpo full i1

blossom v3 baichuan2 7b i1

blossom v1 baichuan 7b i1

👁 01.AI
Yi 1.5 6B

C45

6B15.1 GB84 tok/s4K ctx

dense

Yi 1.5 6B Chat

stablelm 2 1 6b chat imatrix

👁 BigCode
StarCoder2 7B

C44

7B15.3 GB98 tok/s16K ctx

dense

HelpingAI2 6B

C44

HelpingAI2 6B i1

C44

5B14.1 GB70 tok/s2251K ctx

👁 NVIDIA
Nemotron Mini 4B

C44

4B14.9 GB56 tok/s4K ctx

dense

HelpingAI2.5 5B i1

C44

4B13.4 GB56 tok/s2835K ctx

👁 Unsloth
Qwen3.5 4B

C44

4B13.4 GB56 tok/s2835K ctx

dense

gemma 3 4b it

C43

4B13.4 GB56 tok/s2835K ctx

gemma 3 4b it

C43

3B13.0 GB42 tok/s3793K ctx

👁 Alibaba
Qwen 3 0.6B

C43

0.6B11.7 GB8 tok/s33K ctx

dense

👁 Meta
Llama 3.2 1B

C43

1B11.6 GB14 tok/s128K ctx

dense

Llama 3.2 3B Instruct

C43

👁 Qwen
Qwen2.5 3B Instruct

C42

3B12.7 GB42 tok/s3808K ctx

dense

Llama 3.2 3B Instruct

Ministral 3 3B Instruct 2512

HELVETE 3B

Hermes 3 Llama 3.2 3B

stablelm zephyr 3b

StarCoder2 3B

AI21 Jamba Reasoning 3B

stablelm 3b 4e1t

ai21labs AI21 Jamba Reasoning 3B

ai21labs AI21 Jamba2 3B

HelpingAI 3B hindi i1

AI21 Jamba2 3B

HelpingAI 3B hindi

AI21 Jamba2 3B i1

gemma 2 2b it

2B12.4 GB28 tok/s5725K ctx

EXAONE 3.5 2.4B Instruct

2.4B12.2 GB34 tok/s4781K ctx

👁 Google
gemma 2b

C41

2B12.0 GB28 tok/s5754K ctx

dense

Gemmasutra Mini 2B v1

2B12.0 GB28 tok/s5754K ctx

gemma 2 2b it

2B12.0 GB28 tok/s5754K ctx

1.5B11.6 GB21 tok/s7699K ctx

👁 Alibaba
Qwen 3.5 0.6B

C41

0.6B11.7 GB8 tok/s131K ctx

dense

👁 BigCode
StarCoder2 3B

C41

3B12.8 GB42 tok/s16K ctx

dense

F Tier -- Can't Run (44 models)

Too heavy for this hardware.

👁 Qwen
Qwen2.5 1.5B Instruct

D40

1.5B11.6 GB21 tok/s7699K ctx

dense

👁 Unsloth
DeepSeek R1 Distill Qwen 1.5B

D40

1.5B11.6 GB21 tok/s7699K ctx

dense

Yi Coder 1.5B Chat

D40

1.6B11.7 GB22 tok/s7213K ctx

stablelm 2 zephyr 1.6b

D40

logos16v2 stablelm2 1.6b i1

D40

1.6B11.7 GB22 tok/s7213K ctx

Yi Coder 1.5B

1.5B11.6 GB21 tok/s7699K ctx

👁 Unsloth
Falcon H1 1.5B Instruct

D39

1.5B11.6 GB21 tok/s7699K ctx

dense

TinyLlama 1.1B Chat v1.0

1.1B11.3 GB15 tok/s10529K ctx

👁 Alibaba
Qwen 2.5 0.5B

D39

0.5B11.0 GB7 tok/s131K ctx

dense

EXAONE 4.0 1.2B

1.2B11.4 GB17 tok/s9645K ctx

Llama 3.2 1B Instruct Q8 0

1B11.4 GB14 tok/s11562K ctx

TinyLlama 1.1B Chat v0.3

1.1B11.3 GB15 tok/s10529K ctx

👁 TinyLlama
TinyLlama 1.1B Chat v0.6

D39

1.1B11.3 GB15 tok/s10529K ctx

dense

Llama 3.2 1B Instruct

1B11.2 GB14 tok/s11590K ctx

gemma 3 1b it

1B11.2 GB14 tok/s11590K ctx

TinyLlama 1.1B Chat v1.0 imatrix

1.1B11.3 GB15 tok/s10529K ctx

Falcon3 1B Instruct abliterated

1B11.2 GB14 tok/s11590K ctx

0.5B11.0 GB7 tok/s13614K ctx

SmolVLM 500M Instruct

D36

0.3B10.8 GB4 tok/s13641K ctx

embeddinggemma 300M

D35

granite embedding 107m multilingual

D33

0.11B10.7 GB2 tok/s13670K ctx

Falcon H1 Tiny 90M Instruct

D33

0.09B10.7 GB2 tok/s13671K ctx

230B154.6 GB29 tok/s4K ctx

👁 Alibaba
Qwen 3 235B A22B

235B156.7 GB25 tok/s4K ctx

moe

👁 Alibaba
Qwen3-Coder 480B A35B Instruct

480B306.2 GB5 tok/s4K ctx

moe

👁 Alibaba
Qwen 3.5 397B A17B

397B255.5 GB7 tok/s4K ctx

moe

👁 Mistral
Mistral Large 3

675B427.6 GB4 tok/s4K ctx

+1moe

👁 Z.ai
GLM-5

744B483.4 GB3 tok/s4K ctx

moe

👁 Z.ai
GLM-5.1

754B489.5 GB3 tok/s4K ctx

moe

👁 Moonshot AI
Kimi K2.5

1000B627.9 GB3 tok/s4K ctx

moe

👁 Moonshot AI
Kimi K2.6

1000B627.9 GB3 tok/s4K ctx

+1moe

👁 DeepSeek
DeepSeek V3 671B

671B479.4 GB4 tok/s4K ctx

moe

👁 DeepSeek
DeepSeek V3.1 671B

671B479.4 GB4 tok/s4K ctx

moe

👁 DeepSeek
DeepSeek V4 Pro

1600B874.4 GB2 tok/s4K ctx

moe

👁 DeepSeek
DeepSeek V4 Flash

284B169.8 GB23 tok/s4K ctx

moe

👁 DeepSeek
DeepSeek R1 671B

671B479.4 GB4 tok/s4K ctx

moe

👁 DeepSeek
DeepSeek Coder V2 236B

236B213.1 GB14 tok/s4K ctx

moe

👁 DeepSeek
DeepSeek V2.5 236B

236B213.1 GB14 tok/s4K ctx

moe

👁 Meta
Llama 4 Maverick 17B 128E

400B257.4 GB7 tok/s4K ctx

moe

👁 Meta
Llama 3.1 405B

405B265.2 GB2 tok/s4K ctx

dense

👁 DeepSeek
DeepSeek V3.2

671B420.3 GB4 tok/s4K ctx

moe

MiniMax M2.7

moe

K EXAONE 236B A23B

236B182.1 GB7 tok/s4K ctx

235B181.4 GB7 tok/s4K ctx

Baichuan M3 235B

235B181.4 GB7 tok/s4K ctx

Baichuan M3 235B i1

👁 Unsloth
Qwen3.5 397B A17B

397B299.2 GB2 tok/s4K ctx

dense

Compare with...

See How Other Hardware Stacks Up

Check tier lists for similar GPUs and Apple Silicon configs

RTX PRO 6000 Blackwell Workstation Edition 96GB

96 GB VRAM

RTX PRO 6000 Blackwell Server Edition 96GB

96 GB VRAM

NVIDIA H20 96GB

96 GB VRAM

NVIDIA A100 80GB

80 GB VRAM