VOOZH about

URL: https://willitrunai.com/my-results/instinct-mi325x-256gb

⇱ AMD Instinct MI325X 256GB AI Tier List: 341 models ranked | WillItRunAI


AI Compatibility Tier List

AMD Instinct MI325X 256GB

374 AI models ranked by how well they run on your 256 GB VRAM hardware. Best match: DeepSeek V4 Flash (score: 99).

374

Models tested

341

Can run

256 GB

VRAM

94

Best tok/s

Tier Distribution

S 28
A 55
B 66
C 192
F 33
S: 28A: 55B: 66C: 192F: 33
S

S Tier -- Perfect Fit (28 models)

These models run flawlessly with headroom to spare.

👁 DeepSeek
DeepSeek V4 Flash
S99
284B185.8 GB94 tok/s872K ctx
moe
👁 Alibaba
Qwen 3 235B A22B
S95
235B172.7 GB89 tok/s131K ctx
moe
👁 Alibaba
Qwen 3.5 122B A10B
S93
122B103.4 GB176 tok/s131K ctx
moe
👁 Mistral
Devstral 2 123B Instruct
S93
123B106.9 GB64 tok/s256K ctx
dense
MiniMax M2.7
S93
230B170.6 GB101 tok/s205K ctx
moe
👁 Mistral
Mistral Small 4 119B
S91
119B104.5 GB191 tok/s256K ctx
moe
👁 OpenAI
GPT-OSS 120B
S90
117B102.8 GB67 tok/s131K ctx
dense
👁 Cohere
Command A 111B
S90
111B98.1 GB71 tok/s262K ctx
dense
👁 DeepSeek
DeepSeek Coder V2 236B
S89
236B229.1 GB82 tok/s23K ctx
moe
👁 Mistral AI
Pixtral Large 124B
S89
124B107.5 GB63 tok/s131K ctx
dense
👁 Alibaba
Qwen3-Coder 30B A3B Instruct
S89
30.5B46.6 GB662 tok/s256K ctx
moe
👁 Alibaba
Qwen 2.5 VL 72B
S89
72B75.3 GB109 tok/s33K ctx
dense
👁 Alibaba
Qwen 3.6 35B A3B
S88
35B52.0 GB557 tok/s262K ctx
+1moe
👁 Alibaba
Qwen 3.5 27B
S88
27B46.1 GB287 tok/s131K ctx
dense
👁 Alibaba
Qwen3-Coder-Next
S88
80B76.8 GB296 tok/s256K ctx
moe
👁 Alibaba
Qwen 3.6 27B
S88
27B43.9 GB179 tok/s262K ctx
+1dense
👁 Alibaba
Qwen3-VL 30B A3B Instruct
S88
30B46.3 GB685 tok/s256K ctx
moe
👁 DeepSeek
DeepSeek V2.5 236B
S88
236B229.1 GB82 tok/s23K ctx
moe
👁 Alibaba
Qwen 3.5 35B A3B
S88
35B49.3 GB605 tok/s131K ctx
moe
👁 Alibaba
Qwen 3 32B
S87
32B49.9 GB244 tok/s131K ctx
dense
👁 Mistral
Magistral Small 2507
S87
24B43.6 GB322 tok/s131K ctx
dense
👁 Mistral
Devstral Small 2 24B Instruct
S87
24B43.6 GB322 tok/s256K ctx
dense
👁 Mistral
Leanstral 119B A6B
S87
119B107.9 GB176 tok/s256K ctx
moe
👁 Alibaba
Qwen 3.5 9B
S86
9B34.2 GB126 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 30B A3B
S86
30.5B46.6 GB662 tok/s131K ctx
moe
👁 NVIDIA
Nemotron 3 Nano 30B
S86
30B47.2 GB257 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 14B
S86
14B37.5 GB196 tok/s131K ctx
dense
👁 Mistral
Devstral Small 1.1
S85
24B43.6 GB322 tok/s131K ctx
dense
A

A Tier -- Great Fit (55 models)

Excellent performance with comfortable memory margins.

👁 Microsoft
Phi-4-reasoning-plus 14B
A85
14.7B38.5 GB206 tok/s33K ctx
dense
👁 Alibaba
Qwen 3 8B
A85
8B33.6 GB112 tok/s131K ctx
dense
👁 Alibaba
Qwen 3.5 397B A17B
A84
397B271.5 GB39 tok/s4K ctx
moe
👁 OpenAI
GPT-OSS 20B
A84
21B41.8 GB841 tok/s128K ctx
moe
👁 Google
Gemma 4 31B
A84
30.7B59.9 GB153 tok/s230K ctx
dense
👁 NVIDIA
Nemotron Cascade 2 30B A3B
A84
30B47.7 GB677 tok/s262K ctx
moe
👁 Alibaba
Qwen 3.5 4B
A83
4B31.1 GB56 tok/s131K ctx
dense
👁 Meta
Llama 3.3 70B
A82
70B74.1 GB112 tok/s128K ctx
dense
👁 LG AI
EXAONE 4.0 32B
A81
32B49.9 GB242 tok/s131K ctx
dense
👁 Google
Gemma 4 26B A4B
A81
25.2B45.5 GB711 tok/s256K ctx
moe
👁 Mistral
Ministral 3 14B
A80
14B37.5 GB196 tok/s262K ctx
multimodal
👁 Microsoft
Phi-4 Mini Reasoning 4B
A80
3.8B30.3 GB53 tok/s131K ctx
dense
👁 NVIDIA
Nemotron Nano 8B
A80
8B33.3 GB112 tok/s131K ctx
dense
👁 Mistral
Codestral 2 25.08
A80
22B42.4 GB308 tok/s256K ctx
dense
👁 Alibaba
Qwen 2.5 32B
A80
32B49.9 GB242 tok/s131K ctx
dense
👁 Mistral
Mistral Small 3.2 24B
A79
24B43.6 GB322 tok/s131K ctx
vision
👁 Alibaba
Qwen 2.5 72B
A79
72B75.3 GB109 tok/s131K ctx
dense
👁 Meta
Llama 3.1 70B
A79
70B74.1 GB112 tok/s128K ctx
dense
👁 Moonshot AI
Kimi Linear 48B A3B
A79
48B56.7 GB150 tok/s1049K ctx
linear
👁 Google
Gemma 3 27B
A79
27B54.2 GB174 tok/s131K ctx
dense
A79
24B43.6 GB322 tok/s131K ctx
dense
A78
19B40.5 GB266 tok/s8K ctx
dense
👁 Allen AI
OLMo 2 32B
A78
32B49.9 GB242 tok/s4K ctx
dense
👁 IBM
Granite 4.1 30B
A78
30B48.7 GB257 tok/s131K ctx
dense
👁 Meta
Llama 4 Scout 17B 16E
A78
109B95.9 GB168 tok/s890K ctx
moe
👁 Mistral
Mistral Small 24B
A77
24B43.6 GB322 tok/s33K ctx
dense
👁 InternLM
InternVL2 8B
A77
8B33.3 GB112 tok/s8K ctx
dense
👁 Mistral AI
Mistral Small 3.1 24B
A77
24B43.6 GB322 tok/s131K ctx
dense
A77
14B38.1 GB196 tok/s16K ctx
dense
👁 Alibaba
Qwen 2.5 14B
A76
14B38.0 GB196 tok/s131K ctx
dense
👁 Meta
Llama 4 Maverick 17B 128E
A75
400B273.4 GB38 tok/s4K ctx
moe
👁 Mistral
Ministral 3 8B
A75
8B33.6 GB112 tok/s262K ctx
multimodal
👁 OpenBMB
MiniCPM-V 2.6 8B
A75
8B33.3 GB112 tok/s2K ctx
dense
👁 IBM
Granite Code 20B
A75
20B41.9 GB280 tok/s8K ctx
dense
👁 Google
Gemma 3 12B
A75
12B38.7 GB168 tok/s131K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 70B
A74
70B74.1 GB112 tok/s131K ctx
dense
👁 Alibaba
Qwen 3 4B
A74
4B31.1 GB56 tok/s33K ctx
dense
👁 Defog
SQLCoder 7B
A74
7B32.7 GB98 tok/s8K ctx
dense
👁 NVIDIA
Nemotron Nano 9B v2
A74
9B34.4 GB126 tok/s131K ctx
dense
👁 Alibaba
Qwen 2.5 VL 7B
A74
7B31.6 GB98 tok/s33K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 32B
A74
32B49.9 GB242 tok/s131K ctx
dense
👁 DeepSeek
DeepSeek Coder V2 16B
A74
16B39.6 GB1069 tok/s131K ctx
moe
👁 Jina AI
Jina Embeddings v3
A73
0.57B29.6 GB8 tok/s8K ctx
dense
👁 IBM
Granite Code 34B
A73
34B50.9 GB229 tok/s8K ctx
dense
👁 BAAI
BGE M3
A73
0.57B28.8 GB8 tok/s8K ctx
dense
👁 Magistral AI
Magistral 7B
A72
7B32.7 GB98 tok/s8K ctx
dense
👁 Tsinghua/Zhipu
CodeGeeX 4 9B
A72
9B32.6 GB126 tok/s131K ctx
dense
👁 Cohere
Command R 35B
A72
35B50.3 GB223 tok/s131K ctx
dense
👁 Google
Gemma 4 E4B
A72
8B32.7 GB112 tok/s128K ctx
dense
👁 AllenAI
OLMo 2 13B
A72
13B36.9 GB182 tok/s33K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 32B
A71
32B49.9 GB242 tok/s33K ctx
dense
👁 BigCode
StarCoder 15B
A71
15B51.9 GB210 tok/s8K ctx
dense
👁 Meta
Llama 3.1 405B
A71
405B281.2 GB12 tok/s4K ctx
dense
👁 Meta
CodeLlama 13B Instruct
A71
13B46.6 GB182 tok/s16K ctx
dense
👁 Alibaba
Qwen 2.5 7B
A70
7B31.6 GB98 tok/s131K ctx
dense
B

B Tier -- Good Fit (66 models)

Solid performance, may need quantization or reduced context.

👁 IBM
Granite Code 8B
B70
8B33.3 GB112 tok/s8K ctx
dense
👁 NVIDIA
Nemotron 70B
B70
70B74.1 GB112 tok/s131K ctx
dense
👁 DevStral AI
DevStral 7B
B70
7B32.7 GB98 tok/s8K ctx
dense
👁 Mistral AI
Codestral Mamba 7B
B70
7B31.3 GB98 tok/s262K ctx
state-space
👁 DeepSeek
DeepSeek R1 Distill 14B
B70
14B38.0 GB196 tok/s33K ctx
dense
👁 Mixedbread AI
mxbai Embed Large
B69
0.34B28.7 GB5 tok/s1K ctx
dense
👁 LLaVA
LLaVA 1.6 13B
B69
13B46.6 GB182 tok/s4K ctx
dense
👁 IBM
Granite 4.1 8B
B69
8B33.8 GB112 tok/s131K ctx
dense
👁 Mistral AI
Pixtral 12B
B69
12B36.3 GB168 tok/s131K ctx
dense
👁 MosaicML
MPT-30B-Instruct
B68
30B71.5 GB207 tok/s8K ctx
dense
👁 Meta
CodeLlama 7B Instruct
B68
7B38.6 GB98 tok/s16K ctx
dense
👁 BigCode
StarCoder 7B
B68
7B38.1 GB98 tok/s8K ctx
dense
👁 Snowflake
Snowflake Arctic Embed L
B68
0.34B28.7 GB5 tok/s1K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 3B
B68
3B30.5 GB42 tok/s131K ctx
dense
👁 TII
Falcon 40B Instruct
B67
40B57.1 GB169 tok/s8K ctx
dense
👁 Nous Research
Nous Dolphin 13B
B67
13B48.1 GB182 tok/s16K ctx
dense
👁 Google
Gemma 4 E2B
B66
5.1B30.1 GB71 tok/s128K ctx
dense
👁 BAAI
BGE Large EN v1.5
B66
0.34B28.7 GB5 tok/s1K ctx
dense
👁 WizardLM
WizardLM 13B
B66
13B46.6 GB182 tok/s8K ctx
dense
👁 LMSYS
Vicuna 13B
B66
13B46.6 GB182 tok/s4K ctx
dense
👁 Cohere
Command R+ 104B
B66
104B93.4 GB75 tok/s131K ctx
dense
👁 Nous Research
Nous Hermes 1.0
B66
9B44.2 GB126 tok/s16K ctx
dense
👁 Google
Gemma 2 27B
B66
27B54.2 GB174 tok/s8K ctx
dense
👁 Zhipu
GLM-4 9B
B66
9B32.6 GB126 tok/s128K ctx
dense
👁 InternLM
InternLM 7B
B66
7B38.6 GB98 tok/s8K ctx
dense
👁 InternLM
InternLM Chat 7B
B66
7B38.6 GB98 tok/s8K ctx
dense
👁 Meta
Llama 3.1 8B
B66
8B33.3 GB112 tok/s128K ctx
dense
B65
7B33.7 GB98 tok/s8K ctx
dense
👁 WizardLM
WizardMath 7B
B65
7B32.7 GB98 tok/s4K ctx
dense
👁 Allen AI
OLMo 2 7B
B65
7B32.7 GB98 tok/s4K ctx
dense
👁 Nomic AI
Nomic Embed Text v1.5
B65
0.14B27.3 GB2 tok/s8K ctx
dense
👁 Mistral
Mixtral 8x22B
B65
141B115.9 GB106 tok/s66K ctx
moe
👁 Alibaba
Qwen 2.5 Coder 7B
B64
7B31.6 GB98 tok/s131K ctx
dense
👁 Mistral
Ministral 3 3B
B64
3B29.1 GB42 tok/s262K ctx
multimodal
👁 Google
Gemma 3 4B
B64
4B31.0 GB56 tok/s128K ctx
dense
👁 Microsoft
Phi 4 Mini 4B
B63
4B30.4 GB56 tok/s128K ctx
dense
👁 LLaVA
LLaVA 1.5 7B
B63
7B38.6 GB98 tok/s4K ctx
dense
👁 Alibaba
Qwen 3.5 2B
B63
2B29.4 GB28 tok/s131K ctx
dense
👁 Mistral
Mixtral 8x7B
B62
47B57.1 GB315 tok/s33K ctx
moe
👁 Baichuan
Baichuan 13B
B62
13B48.1 GB182 tok/s8K ctx
dense
👁 Cerebras
Cerebras-GPT 13B
B62
13B45.6 GB182 tok/s131K ctx
dense
👁 Alibaba
Qwen 2.5 Math 72B
B62
72B75.3 GB109 tok/s4K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 7B
B62
7B31.6 GB98 tok/s33K ctx
dense
👁 MosaicML
MPT-7B-Instruct
B61
7B38.6 GB98 tok/s8K ctx
dense
👁 Baichuan
Baichuan 7B
B61
7B38.6 GB98 tok/s8K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 8B
B61
8B33.3 GB112 tok/s33K ctx
dense
7B32.7 GB98 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 3B
B61
3B30.5 GB42 tok/s131K ctx
dense
👁 TII
Falcon 7B Instruct
B61
7B30.9 GB98 tok/s8K ctx
dense
👁 Microsoft
Phi 3 Mini 3.8B
B60
3.8B34.7 GB53 tok/s128K ctx
dense
👁 Alibaba
Qwen 3 1.7B
B60
1.7B29.2 GB24 tok/s33K ctx
dense
B60
0.02B26.8 GB2 tok/s0K ctx
dense
👁 Google
Gemma 2 9B
B59
9B37.1 GB126 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 14B
B59
14B38.0 GB196 tok/s131K ctx
dense
👁 Meta
Llama 3.2 11B Vision
B59
11B35.2 GB154 tok/s16K ctx
vision
👁 Microsoft
Phi 3.5 Mini 4B
B59
4B34.8 GB56 tok/s128K ctx
dense
👁 IBM
Granite 4.1 3B
B59
3B29.6 GB42 tok/s131K ctx
dense
👁 01.AI
Yi 1.5 34B
B59
34B50.9 GB229 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 1.5B
B58
1.5B27.8 GB21 tok/s33K ctx
dense
👁 IBM
Granite Code 3B
B58
3B30.8 GB42 tok/s8K ctx
dense
👁 DeepSeek
DeepSeek LLM 67B
B58
67B73.2 GB117 tok/s4K ctx
dense
👁 Mistral
Mistral Nemo 12B
B57
12B36.3 GB168 tok/s128K ctx
dense
👁 01.AI
Yi Coder 9B
B57
9B33.5 GB126 tok/s131K ctx
dense
👁 Mistral
Mistral 7B Instruct v0.3
B57
7B32.7 GB98 tok/s8K ctx
dense
👁 Microsoft
Phi 3 Medium 14B
B56
14B38.1 GB196 tok/s128K ctx
dense
👁 Mistral AI
Codestral 22B
B55
22B42.4 GB308 tok/s33K ctx
dense
C

C Tier -- Tight Fit (192 models)

Runs with offloading or significant compromises.

👁 InternLM
InternLM 20B
C55
20B59.2 GB280 tok/s8K ctx
dense
👁 Meta
Llama 3.2 3B
C54
3B30.0 GB42 tok/s128K ctx
dense
👁 Mistral
Ministral 8B
C54
8B33.6 GB112 tok/s131K ctx
dense
K EXAONE 236B A23B
C53
236B198.1 GB30 tok/s49K ctx
dense
Baichuan M3 235B
C53
235B197.4 GB31 tok/s50K ctx
dense
Baichuan M3 235B i1
C53
235B197.4 GB31 tok/s50K ctx
dense
👁 Cohere
Aya Expanse 32B
C51
32B48.5 GB244 tok/s8K ctx
dense
👁 DeepSeek
DeepSeek R1 1.5B
C50
1.5B27.8 GB21 tok/s33K ctx
dense
👁 TinyLlama
TinyLlama 1.1B
C50
1.1B27.5 GB15 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 122B A10B
C50
122B100.6 GB68 tok/s190K ctx
dense
C50
3B30.3 GB42 tok/s128K ctx
dense
👁 IBM
Granite 3.1 8B
C49
8B33.3 GB112 tok/s128K ctx
state-space
Solar Open 100B
C49
100B99.2 GB72 tok/s230K ctx
dense
Solar Open 100B i1
C49
100B99.2 GB72 tok/s230K ctx
dense
👁 01.AI
Yi 1.5 9B
C49
9B33.5 GB126 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 Math 7B
C49
7B31.6 GB98 tok/s4K ctx
dense
Llama 3.3 70B Instruct
C48
70B77.4 GB103 tok/s364K ctx
dense
Solar Open 69B REAP i1
C48
69B76.7 GB104 tok/s371K ctx
dense
stabilityai japanese stablelm instruct beta 70b
C48
70B77.4 GB103 tok/s364K ctx
dense
BaichuanMed OCR 72B i1
C48
72B78.9 GB100 tok/s352K ctx
dense
👁 01.AI
Yi 34B Chat
C48
34B50.9 GB229 tok/s200K ctx
dense
👁 Alibaba
Qwen 2.5 1.5B
C48
1.5B27.8 GB21 tok/s131K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
C47
48B61.4 GB150 tok/s570K ctx
dense
👁 OpenChat
OpenChat 7B
C47
7B32.7 GB98 tok/s8K ctx
dense
👁 Unsloth
Qwen3.5 35B A3B
C47
35B52.0 GB205 tok/s812K ctx
dense
👁 BigCode
StarCoder2 15B
C46
15B38.5 GB210 tok/s16K ctx
dense
Qwen3.5 35B A3B
C46
35B52.0 GB205 tok/s812K ctx
dense
👁 Cohere
Aya Expanse 8B
C46
8B33.3 GB112 tok/s8K ctx
dense
👁 Google
Gemma 3 1B
C46
1B27.5 GB14 tok/s33K ctx
dense
👁 Unsloth
Qwen3.5 27B
C46
27B46.1 GB266 tok/s1077K ctx
dense
Hermes 4.3 36B
C46
36B52.7 GB200 tok/s787K ctx
dense
👁 Unsloth
gemma 3 27b it
C46
27B46.1 GB266 tok/s1077K ctx
dense
EXAONE 4.0 32B
C46
32B49.8 GB224 tok/s896K ctx
dense
gemma 3 27b it
C46
27B46.1 GB266 tok/s1077K ctx
dense
aya expanse 32b heretic MPOA i1
C46
32B49.8 GB224 tok/s896K ctx
dense
Baichuan M2 32B Q4 K M
C46
32B49.8 GB224 tok/s896K ctx
dense
👁 Unsloth
Mistral Small 3.2 24B Instruct 2506
C46
24B44.0 GB299 tok/s1222K ctx
dense
👁 LMSYS
Vicuna 7B
C46
7B38.6 GB98 tok/s4K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C46
24B44.0 GB299 tok/s1222K ctx
dense
baichuan inc Baichuan M2 32B
C46
32B49.8 GB224 tok/s896K ctx
dense
Mistral Small 24B Instruct 2501
C46
24B44.0 GB299 tok/s1222K ctx
dense
👁 Stability AI
StableLM 2 12B
C46
12B47.3 GB168 tok/s4K ctx
dense
mistral small 3.1 24b instruct 2503 hf
C46
24B44.0 GB299 tok/s1222K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
C46
24B44.0 GB299 tok/s1222K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
C45
24B44.0 GB299 tok/s1222K ctx
dense
👁 Teknium
OpenHermes 2.5 7B
C45
7B32.7 GB98 tok/s8K ctx
dense
Codestral 22B v0.1
C45
22B42.5 GB308 tok/s1341K ctx
dense
Codestral 22B v0.1
C45
22B42.5 GB308 tok/s1341K ctx
dense
8B33.3 GB112 tok/s33K ctx
dense
👁 Google
Gemma 2 2B
C45
2B29.3 GB28 tok/s8K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
C45
24B44.0 GB299 tok/s1222K ctx
dense
Codestral 22B v0.1
C45
22B42.5 GB308 tok/s1341K ctx
dense
👁 Hugging Face H4
Zephyr 7B Beta
C45
7B32.7 GB98 tok/s33K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Qwen 14B
C45
14B36.7 GB196 tok/s2155K ctx
dense
Codestral 22B v0.1 IMat
C45
22B42.5 GB308 tok/s1341K ctx
dense
👁 Unsloth
Qwen3.5 9B
C45
9B33.0 GB126 tok/s3398K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
C45
9B33.0 GB126 tok/s3398K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 0.5B
C45
0.5B27.0 GB7 tok/s131K ctx
dense
gemma 3 12b it
C45
12B35.2 GB168 tok/s2528K ctx
dense
Meta Llama 3.1 8B Instruct
C45
8B32.3 GB112 tok/s3834K ctx
dense
Phi 4 reasoning vision 15B
C45
15B37.4 GB210 tok/s2006K ctx
dense
internlm2 5 20b chat
C45
20B41.0 GB280 tok/s1483K ctx
dense
👁 UC Berkeley
Starling LM 7B
C45
7B32.7 GB98 tok/s8K ctx
dense
Codestral 22B v0.1 i1
C45
22B42.5 GB308 tok/s1341K ctx
dense
StarCoder2 15B
C45
15B37.4 GB210 tok/s2006K ctx
dense
Codestral 21B Pruned i1
C45
21B41.8 GB294 tok/s1409K ctx
dense
llava llama 3 8b v1 1
C45
8B32.3 GB112 tok/s3834K ctx
dense
internlm2 limarp chat 20b
C45
20B41.0 GB280 tok/s1483K ctx
dense
Qwen3.5 9B
C45
9B33.0 GB126 tok/s3398K ctx
dense
👁 Unsloth
DeepSeek R1 0528 Qwen3 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
NousResearch Hermes 4 14B
C45
14B36.7 GB196 tok/s2155K ctx
dense
Llama 2 7B Chat
C45
7B31.6 GB98 tok/s4393K ctx
dense
internlm2 math plus 20b i1
C45
20B41.0 GB280 tok/s1483K ctx
dense
Meta Llama 3 8B Instruct
C45
8B32.3 GB112 tok/s3834K ctx
dense
Mistral 7B Instruct v0.2
C45
7B31.6 GB98 tok/s4393K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Llama 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
Dolphin3.0 Llama3.1 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
Llama 3 8B Instruct 32k v0.1
C45
8B32.3 GB112 tok/s3834K ctx
dense
Mistral 7B Instruct v0.3
C45
7B31.6 GB98 tok/s4393K ctx
dense
Meta Llama 3.1 8B Instruct
C45
8B32.3 GB112 tok/s3834K ctx
dense
vntl llama3 8b v2
C45
8B32.3 GB112 tok/s3834K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
Codestral RAG 19B Pruned i1
C45
19B40.3 GB266 tok/s1566K ctx
dense
Yi Coder 9B Chat
C45
9B33.0 GB126 tok/s3398K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C45
10.7B34.3 GB150 tok/s2845K ctx
dense
starcoder2 15b instruct v0.1
C45
15B37.4 GB210 tok/s2006K ctx
dense
glm 4 9b chat 1m
C45
9B33.0 GB126 tok/s3398K ctx
dense
starcoder2 15b i1
C45
15B37.4 GB210 tok/s2006K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C45
10.7B34.3 GB150 tok/s2845K ctx
dense
DeepSeek R1 0528 Qwen3 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
starcoder2 15b instruct v0.1
C45
15B37.4 GB210 tok/s2006K ctx
dense
Qwen3 8B DeepSeek v3.2 Speciale Distill
C45
8B32.3 GB112 tok/s3834K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
C45
14B36.7 GB196 tok/s2155K ctx
dense
Hermes 3 Llama 3.1 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
internlm JanusCoder 14B
C45
14B36.7 GB196 tok/s2155K ctx
dense
HelpingAI 15B i1
C45
15B37.4 GB210 tok/s2006K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
C45
14B36.7 GB196 tok/s2155K ctx
dense
Yi 1.5 9B Chat
C45
9B33.0 GB126 tok/s3398K ctx
dense
Hermes 2 Pro Llama 3 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
Mistral 7B Instruct v0.3
C45
7B31.6 GB98 tok/s4393K ctx
dense
Dolphin3.0 Llama3.1 8B
C45
8B32.3 GB112 tok/s3834K ctx
dense
Hermes 2 Pro Mistral 7B
C45
7B31.6 GB98 tok/s4393K ctx
dense
granite 8b code instruct 4k
C44
8B32.3 GB112 tok/s3834K ctx
dense
zephyr 7B beta
C44
7B31.6 GB98 tok/s4393K ctx
dense
dolphin 2.9.4 llama3.1 8b
C44
8B32.3 GB112 tok/s3834K ctx
dense
Nous Hermes 2 Mistral 7B DPO
C44
7B31.6 GB98 tok/s4393K ctx
dense
Falcon H1R 7B
C44
7B31.6 GB98 tok/s4393K ctx
dense
solar finalised finetuned Model 10.7B i1
C44
10.7B34.3 GB150 tok/s2845K ctx
dense
openchat 3.6 8b 20240522 IMat
C44
8B32.3 GB112 tok/s3834K ctx
dense
falcon mamba 7b instruct Q4 K M
C44
7B31.6 GB98 tok/s4393K ctx
dense
EXAONE 3.5 7.8B Instruct
C44
7.8B32.2 GB109 tok/s3934K ctx
dense
aya expanse 8b
C44
8B32.3 GB112 tok/s3834K ctx
dense
zephyr 7B alpha
C44
7B31.6 GB98 tok/s4393K ctx
dense
speechless zephyr code functionary 7b
C44
7B31.6 GB98 tok/s4393K ctx
dense
SOLAR 10.7B v1.0
C44
10.7B34.3 GB150 tok/s2845K ctx
dense
dolphin v2 8b abliterated i1
C44
8B32.3 GB112 tok/s3834K ctx
dense
EXAONE 3.5 7.8B Instruct
C44
7.8B32.2 GB109 tok/s3934K ctx
dense
StarCoder2 7B
C44
7B31.6 GB98 tok/s4393K ctx
dense
EXAONE 3.5 7.8B Instruct i1
C44
7.8B32.2 GB109 tok/s3934K ctx
dense
japanese stablelm instruct gamma 7B
C44
7B31.6 GB98 tok/s4393K ctx
dense
Yi 9B Coder i1
C44
9B33.0 GB126 tok/s3398K ctx
dense
aya expanse 8b orthogonal heretic i1
C44
8B32.3 GB112 tok/s3834K ctx
dense
exaone 3.0 7.8b it
C44
7.8B32.2 GB109 tok/s3934K ctx
dense
Yi 1.5 6B Chat
C44
6B30.9 GB84 tok/s5139K ctx
dense
Falcon H1 7B Instruct
C44
7B31.6 GB98 tok/s4393K ctx
dense
👁 Unsloth
Falcon H1R 7B
C44
7B31.6 GB98 tok/s4393K ctx
dense
HelpingAI2 9B
C44
9B33.0 GB126 tok/s3398K ctx
dense
stablelm 2 zephyr 1 6b
C44
6B30.9 GB84 tok/s5139K ctx
dense
aya expanse 8b orthogonal heretic
C44
8B32.3 GB112 tok/s3834K ctx
dense
HelpingAI2.5 10B i1
C44
10B33.8 GB140 tok/s3050K ctx
dense
Helply 10.2b chat i1
C44
10.2B33.9 GB143 tok/s2989K ctx
dense
HelpingAI2 9B i1
C44
9B33.0 GB126 tok/s3398K ctx
dense
Mamba Codestral 7B v0.1
C44
7B31.6 GB98 tok/s4393K ctx
dense
stabilityai japanese stablelm base gamma 7b
C44
7B31.6 GB98 tok/s4393K ctx
dense
HelpingAI 9B 200k i1
C44
9B33.0 GB126 tok/s3398K ctx
dense
baichuan2 7b chat
C44
7B31.6 GB98 tok/s4393K ctx
dense
DiscoPOP zephyr 7b gemma
C44
7B31.6 GB98 tok/s4393K ctx
dense
HelpingAI 9B i1
C44
9B33.0 GB126 tok/s3398K ctx
dense
internlm2 5 1 8b chat i1
C44
8B32.3 GB112 tok/s3834K ctx
dense
internlm3 8b instruct abliterated i1
C44
8B32.3 GB112 tok/s3834K ctx
dense
starcoder2 7b
C44
7B31.6 GB98 tok/s4393K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
C44
7B31.6 GB98 tok/s4393K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
internlm2 math plus 7b IMat
C44
7B31.6 GB98 tok/s4393K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
Yi 1.5 6B Chat
C44
6B30.9 GB84 tok/s5139K ctx
dense
internlm2 5 7b chat i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
👁 01.AI
Yi 1.5 6B
C44
6B31.1 GB84 tok/s4K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
C44
7B31.6 GB98 tok/s4393K ctx
dense
stablelm 2 1 6b chat imatrix
C44
6B30.9 GB84 tok/s5139K ctx
dense
MD Judge v0 2 internlm2 7b i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
C44
7B31.6 GB98 tok/s4393K ctx
dense
jointpreferences mistral 7b sft helpful
C44
7B31.6 GB98 tok/s4393K ctx
dense
zephyr 7b dpo full i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
blossom v3 baichuan2 7b i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
blossom v1 baichuan 7b i1
C44
7B31.6 GB98 tok/s4393K ctx
dense
👁 Intel
Neural Chat 7B
C44
7B32.7 GB98 tok/s8K ctx
dense
HelpingAI2 6B
C44
6B30.9 GB84 tok/s5139K ctx
dense
👁 BigCode
StarCoder2 7B
C44
7B31.3 GB98 tok/s16K ctx
dense
HelpingAI2 6B i1
C44
6B30.9 GB84 tok/s5139K ctx
dense
👁 NVIDIA
Nemotron Mini 4B
C44
4B30.9 GB56 tok/s4K ctx
dense
👁 DeepSeek
DeepSeek LLM 7B
C43
7B38.1 GB98 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 4B
C43
4B29.4 GB56 tok/s7750K ctx
dense
HelpingAI2.5 5B i1
C43
5B30.1 GB70 tok/s6184K ctx
dense
gemma 3 4b it
C43
4B29.4 GB56 tok/s7750K ctx
dense
gemma 3 4b it
C43
4B29.4 GB56 tok/s7750K ctx
dense
👁 Alibaba
Qwen 3 0.6B
C43
0.6B27.7 GB8 tok/s33K ctx
dense
👁 Meta
Llama 3.2 1B
C43
1B27.6 GB14 tok/s128K ctx
dense
Llama 3.2 3B Instruct
C42
3B29.0 GB42 tok/s10346K ctx
dense
👁 Qwen
Qwen2.5 3B Instruct
C42
3B28.7 GB42 tok/s10362K ctx
dense
Llama 3.2 3B Instruct
C42
3B28.7 GB42 tok/s10362K ctx
dense
Ministral 3 3B Instruct 2512
C42
3B28.7 GB42 tok/s10362K ctx
dense
HELVETE 3B
C42
3B28.7 GB42 tok/s10362K ctx
dense
Hermes 3 Llama 3.2 3B
C42
3B28.7 GB42 tok/s10362K ctx
dense
stablelm zephyr 3b
C42
3B28.7 GB42 tok/s10362K ctx
dense
StarCoder2 3B
C42
3B28.7 GB42 tok/s10362K ctx
dense
AI21 Jamba Reasoning 3B
C41
3B28.7 GB42 tok/s10362K ctx
dense
stablelm 3b 4e1t
C41
3B28.7 GB42 tok/s10362K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C41
3B28.7 GB42 tok/s10362K ctx
dense
ai21labs AI21 Jamba2 3B
C41
3B28.7 GB42 tok/s10362K ctx
dense
HelpingAI 3B hindi i1
C41
3B28.7 GB42 tok/s10362K ctx
dense
AI21 Jamba2 3B
C41
3B28.7 GB42 tok/s10362K ctx
dense
HelpingAI 3B hindi
C41
3B28.7 GB42 tok/s10362K ctx
dense
AI21 Jamba2 3B i1
C41
3B28.7 GB42 tok/s10362K ctx
dense
EXAONE 3.5 2.4B Instruct
C41
2.4B28.2 GB34 tok/s12973K ctx
dense
gemma 2 2b it
C41
2B28.4 GB28 tok/s15555K ctx
dense
👁 Google
gemma 2b
C41
2B28.0 GB28 tok/s15584K ctx
dense
Gemmasutra Mini 2B v1
C41
2B28.0 GB28 tok/s15584K ctx
dense
👁 Alibaba
Qwen 3.5 0.6B
C41
0.6B27.7 GB8 tok/s131K ctx
dense
gemma 2 2b it
C40
2B28.0 GB28 tok/s15584K ctx
dense
👁 BigCode
StarCoder2 3B
C40
3B28.8 GB42 tok/s16K ctx
dense
F

F Tier -- Can't Run (33 models)

Too heavy for this hardware.

👁 Qwen
Qwen2.5 1.5B Instruct
D40
1.5B27.6 GB21 tok/s20806K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Qwen 1.5B
D40
1.5B27.6 GB21 tok/s20806K ctx
dense
Yi Coder 1.5B Chat
D40
1.5B27.6 GB21 tok/s20806K ctx
dense
stablelm 2 zephyr 1.6b
D39
1.6B27.7 GB22 tok/s19501K ctx
dense
logos16v2 stablelm2 1.6b i1
D39
1.6B27.7 GB22 tok/s19501K ctx
dense
Yi Coder 1.5B
D39
1.5B27.6 GB21 tok/s20806K ctx
dense
👁 Unsloth
Falcon H1 1.5B Instruct
D39
1.5B27.6 GB21 tok/s20806K ctx
dense
TinyLlama 1.1B Chat v1.0
D39
1.1B27.3 GB15 tok/s28403K ctx
dense
👁 Alibaba
Qwen 2.5 0.5B
D39
0.5B27.0 GB7 tok/s131K ctx
dense
EXAONE 4.0 1.2B
D39
1.2B27.4 GB17 tok/s26029K ctx
dense
Llama 3.2 1B Instruct Q8 0
D38
1B27.4 GB14 tok/s31222K ctx
dense
TinyLlama 1.1B Chat v0.3
D38
1.1B27.3 GB15 tok/s28403K ctx
dense
👁 TinyLlama
TinyLlama 1.1B Chat v0.6
D38
1.1B27.3 GB15 tok/s28403K ctx
dense
Llama 3.2 1B Instruct
D38
1B27.2 GB14 tok/s31251K ctx
dense
gemma 3 1b it
D38
1B27.2 GB14 tok/s31251K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
D38
1.1B27.3 GB15 tok/s28403K ctx
dense
Falcon3 1B Instruct abliterated
D38
1B27.2 GB14 tok/s31251K ctx
dense
SmolVLM 500M Instruct
D36
0.5B27.0 GB7 tok/s36654K ctx
dense
embeddinggemma 300M
D35
0.3B26.8 GB4 tok/s36681K ctx
dense
granite embedding 107m multilingual
D33
0.11B26.7 GB2 tok/s36710K ctx
dense
Falcon H1 Tiny 90M Instruct
D32
0.09B26.7 GB2 tok/s36711K ctx
dense
👁 Alibaba
Qwen3-Coder 480B A35B Instruct
F0
480B322.2 GB21 tok/s4K ctx
moe
👁 Mistral
Mistral Large 3
F0
675B443.6 GB8 tok/s4K ctx
+1moe
F0
744B499.4 GB5 tok/s4K ctx
moe
👁 Z.ai
GLM-5.1
F0
754B505.5 GB5 tok/s4K ctx
moe
F0
1000B643.9 GB4 tok/s4K ctx
moe
F0
1000B643.9 GB4 tok/s4K ctx
+1moe
👁 DeepSeek
DeepSeek V3 671B
F0
671B495.4 GB6 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V3.1 671B
F0
671B495.4 GB6 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V4 Pro
F0
1600B890.4 GB3 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek R1 671B
F0
671B495.4 GB6 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V3.2
F0
671B436.3 GB8 tok/s4K ctx
moe
👁 Unsloth
Qwen3.5 397B A17B
F0
397B315.2 GB9 tok/s4K ctx
dense

Compare with...

See How Other Hardware Stacks Up

Check tier lists for similar GPUs and Apple Silicon configs

AMD Instinct MI350X 288GB

288 GB VRAM

View tier list

AMD Instinct MI300X 192GB

192 GB VRAM

View tier list

NVIDIA GB200 192GB

192 GB VRAM

View tier list

B100 192GB

192 GB VRAM

View tier list
Full AMD Instinct MI325X 256GB pageCompare this hardware