VOOZH about

URL: https://willitrunai.com/my-results/rtx-4070-12gb

⇱ RTX 4070 12GB AI Tier List: 232 models ranked | WillItRunAI


AI Compatibility Tier List

RTX 4070 12GB

374 AI models ranked by how well they run on your 12 GB VRAM hardware. Best match: Qwen 3.5 9B (score: 98).

374

Models tested

232

Can run

12 GB

VRAM

72

Best tok/s

Tier Distribution

S 9
A 37
B 63
C 123
F 142
S: 9A: 37B: 63C: 123F: 142
S

S Tier -- Perfect Fit (9 models)

These models run flawlessly with headroom to spare.

👁 Alibaba
Qwen 3.5 9B
S98
9B9.8 GB72 tok/s32K ctx
dense
👁 Alibaba
Qwen 3 8B
S96
8B9.2 GB80 tok/s37K ctx
dense
👁 Alibaba
Qwen 3.5 4B
S93
4B6.7 GB64 tok/s54K ctx
dense
👁 NVIDIA
Nemotron Nano 8B
S91
8B8.9 GB77 tok/s41K ctx
dense
👁 InternLM
InternVL2 8B
S89
8B8.9 GB77 tok/s8K ctx
dense
👁 Microsoft
Phi-4 Mini Reasoning 4B
S88
3.8B5.9 GB61 tok/s83K ctx
dense
👁 Mistral
Ministral 3 8B
S87
8B9.2 GB77 tok/s37K ctx
multimodal
👁 OpenBMB
MiniCPM-V 2.6 8B
S87
8B8.9 GB77 tok/s2K ctx
dense
👁 Defog
SQLCoder 7B
S86
7B8.3 GB88 tok/s8K ctx
dense
A

A Tier -- Great Fit (37 models)

Excellent performance with comfortable memory margins.

👁 Magistral AI
Magistral 7B
A84
7B8.3 GB88 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 VL 7B
A84
7B7.2 GB88 tok/s33K ctx
dense
👁 Alibaba
Qwen 3 4B
A84
4B6.7 GB64 tok/s33K ctx
dense
👁 Tsinghua/Zhipu
CodeGeeX 4 9B
A83
9B8.2 GB69 tok/s116K ctx
dense
👁 Google
Gemma 4 E4B
A82
8B8.3 GB56 tok/s63K ctx
dense
👁 NVIDIA
Nemotron Nano 9B v2
A82
9B10.0 GB68 tok/s29K ctx
dense
👁 DevStral AI
DevStral 7B
A82
7B8.3 GB88 tok/s8K ctx
dense
👁 IBM
Granite Code 8B
A81
8B8.9 GB77 tok/s8K ctx
dense
👁 Alibaba
Qwen 3 14B
A81
14B13.1 GB34 tok/s9K ctx
dense
👁 Jina AI
Jina Embeddings v3
A81
0.57B5.2 GB9 tok/s8K ctx
dense
👁 IBM
Granite 4.1 8B
A81
8B9.4 GB77 tok/s33K ctx
dense
👁 Alibaba
Qwen 2.5 7B
A81
7B7.2 GB88 tok/s105K ctx
dense
👁 Mistral AI
Codestral Mamba 7B
A79
7B6.9 GB102 tok/s184K ctx
state-space
👁 BAAI
BGE M3
A78
0.57B4.4 GB9 tok/s8K ctx
dense
👁 AllenAI
OLMo 2 13B
A78
13B12.5 GB37 tok/s13K ctx
dense
A77
7B9.3 GB88 tok/s8K ctx
dense
👁 Meta
Llama 3.1 8B
A77
8B8.9 GB77 tok/s41K ctx
dense
👁 WizardLM
WizardMath 7B
A77
7B8.3 GB88 tok/s4K ctx
dense
👁 Allen AI
OLMo 2 7B
A77
7B8.3 GB88 tok/s4K ctx
dense
👁 Zhipu
GLM-4 9B
A77
9B8.2 GB69 tok/s116K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 3B
A76
3B6.1 GB48 tok/s59K ctx
dense
👁 Mistral AI
Pixtral 12B
A76
12B11.9 GB58 tok/s17K ctx
dense
👁 Mistral
Ministral 3 14B
A75
14B13.1 GB31 tok/s9K ctx
multimodal
👁 Mixedbread AI
mxbai Embed Large
A75
0.34B4.3 GB5 tok/s1K ctx
dense
👁 Google
Gemma 4 E2B
A75
5.1B5.7 GB82 tok/s128K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 7B
A75
7B7.2 GB88 tok/s105K ctx
dense
👁 Snowflake
Snowflake Arctic Embed L
A74
0.34B4.3 GB5 tok/s1K ctx
dense
👁 Google
Gemma 3 4B
A74
4B6.6 GB64 tok/s58K ctx
dense
👁 Microsoft
Phi-4-reasoning-plus 14B
A73
14.7B14.1 GB25 tok/s5K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 8B
A73
8B8.9 GB77 tok/s33K ctx
dense
7B8.3 GB88 tok/s4K ctx
dense
👁 BAAI
BGE Large EN v1.5
A72
0.34B4.3 GB5 tok/s1K ctx
dense
👁 Microsoft
Phi 4 Mini 4B
A72
4B6.0 GB64 tok/s81K ctx
dense
A72
14B13.7 GB28 tok/s7K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 7B
A72
7B7.2 GB88 tok/s33K ctx
dense
👁 Alibaba
Qwen 2.5 14B
A71
14B13.6 GB29 tok/s7K ctx
dense
👁 Mistral
Ministral 3 3B
A71
3B4.7 GB48 tok/s176K ctx
multimodal
B

B Tier -- Good Fit (63 models)

Solid performance, may need quantization or reduced context.

👁 Alibaba
Qwen 3.5 2B
B70
2B5.0 GB32 tok/s81K ctx
dense
👁 Microsoft
Phi 3 Mini 3.8B
B70
3.8B10.3 GB61 tok/s21K ctx
dense
👁 TII
Falcon 7B Instruct
B70
7B6.5 GB90 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 3B
B70
3B6.1 GB48 tok/s59K ctx
dense
👁 Mistral
Mistral 7B Instruct v0.3
B69
7B8.3 GB88 tok/s8K ctx
dense
👁 Microsoft
Phi 3.5 Mini 4B
B68
4B10.4 GB64 tok/s20K ctx
dense
👁 Nomic AI
Nomic Embed Text v1.5
B68
0.14B2.9 GB2 tok/s8K ctx
dense
👁 01.AI
Yi Coder 9B
B68
9B9.1 GB69 tok/s48K ctx
dense
👁 IBM
Granite Code 3B
B67
3B6.4 GB48 tok/s8K ctx
dense
👁 Meta
Llama 3.2 11B Vision
B67
11B10.8 GB64 tok/s16K ctx
vision
👁 Alibaba
Qwen 3 1.7B
B66
1.7B4.8 GB27 tok/s33K ctx
dense
👁 IBM
Granite 4.1 3B
B66
3B5.2 GB48 tok/s106K ctx
dense
👁 Mistral
Ministral 8B
B66
8B9.2 GB77 tok/s37K ctx
dense
👁 BigCode
StarCoder 7B
B65
7B13.7 GB46 tok/s8K ctx
dense
👁 Mistral
Mistral Nemo 12B
B65
12B11.9 GB58 tok/s17K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 14B
B65
14B13.6 GB29 tok/s7K ctx
dense
👁 Google
Gemma 3 12B
B63
12B14.3 GB22 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 1.5B
B62
1.5B3.4 GB24 tok/s33K ctx
dense
👁 Meta
Llama 3.2 3B
B62
3B5.6 GB48 tok/s76K ctx
dense
B62
0.02B2.4 GB2 tok/s0K ctx
dense
👁 IBM
Granite 3.1 8B
B61
8B8.9 GB96 tok/s41K ctx
state-space
👁 01.AI
Yi 1.5 9B
B60
9B9.1 GB69 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 Math 7B
B59
7B7.2 GB88 tok/s4K ctx
dense
👁 Meta
CodeLlama 7B Instruct
B59
7B14.2 GB43 tok/s12K ctx
dense
👁 OpenChat
OpenChat 7B
B59
7B8.3 GB88 tok/s8K ctx
dense
B58
3B5.9 GB48 tok/s66K ctx
dense
👁 Cohere
Aya Expanse 8B
B58
8B8.9 GB77 tok/s8K ctx
dense
👁 Teknium
OpenHermes 2.5 7B
B57
7B8.3 GB88 tok/s8K ctx
dense
👁 Hugging Face H4
Zephyr 7B Beta
B57
7B8.3 GB88 tok/s33K ctx
dense
8B8.9 GB77 tok/s33K ctx
dense
👁 UC Berkeley
Starling LM 7B
B57
7B8.3 GB88 tok/s8K ctx
dense
👁 Unsloth
Qwen3.5 9B
B56
9B8.6 GB63 tok/s67K ctx
dense
👁 InternLM
InternLM 7B
B56
7B14.2 GB43 tok/s8K ctx
dense
👁 InternLM
InternLM Chat 7B
B56
7B14.2 GB43 tok/s8K ctx
dense
Qwen3.5 9B Uncensored HauhauCS Aggressive
B56
9B8.6 GB63 tok/s67K ctx
dense
Qwen3.5 9B
B56
9B8.6 GB63 tok/s67K ctx
dense
👁 Intel
Neural Chat 7B
B56
7B8.3 GB88 tok/s8K ctx
dense
Yi Coder 9B Chat
B56
9B8.6 GB63 tok/s67K ctx
dense
glm 4 9b chat 1m
B56
9B8.6 GB63 tok/s67K ctx
dense
Yi 1.5 9B Chat
B56
9B8.6 GB63 tok/s67K ctx
dense
Meta Llama 3.1 8B Instruct
B56
8B7.9 GB71 tok/s86K ctx
dense
llava llama 3 8b v1 1
B55
8B7.9 GB71 tok/s86K ctx
dense
👁 Unsloth
DeepSeek R1 0528 Qwen3 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
Meta Llama 3 8B Instruct
B55
8B7.9 GB71 tok/s86K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Llama 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
Dolphin3.0 Llama3.1 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
Llama 3 8B Instruct 32k v0.1
B55
8B7.9 GB71 tok/s86K ctx
dense
Meta Llama 3.1 8B Instruct
B55
8B7.9 GB71 tok/s86K ctx
dense
Yi 9B Coder i1
B55
9B8.6 GB63 tok/s67K ctx
dense
vntl llama3 8b v2
B55
8B7.9 GB71 tok/s86K ctx
dense
DeepSeek R1 0528 Qwen3 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
HelpingAI2 9B
B55
9B8.6 GB63 tok/s67K ctx
dense
DeepSeek R1 0528 Qwen3 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
Qwen3 8B DeepSeek v3.2 Speciale Distill
B55
8B7.9 GB71 tok/s86K ctx
dense
HelpingAI2 9B i1
B55
9B8.6 GB63 tok/s67K ctx
dense
Hermes 3 Llama 3.1 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
HelpingAI 9B 200k i1
B55
9B8.6 GB63 tok/s67K ctx
dense
Helply 10.2b chat i1
B55
10.2B9.5 GB64 tok/s49K ctx
dense
HelpingAI 9B i1
B55
9B8.6 GB63 tok/s67K ctx
dense
Hermes 2 Pro Llama 3 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
Dolphin3.0 Llama3.1 8B
B55
8B7.9 GB71 tok/s86K ctx
dense
granite 8b code instruct 4k
B55
8B7.9 GB71 tok/s86K ctx
dense
dolphin 2.9.4 llama3.1 8b
B55
8B7.9 GB71 tok/s86K ctx
dense
C

C Tier -- Tight Fit (123 models)

Runs with offloading or significant compromises.

openchat 3.6 8b 20240522 IMat
C55
8B7.9 GB71 tok/s86K ctx
dense
aya expanse 8b
C55
8B7.9 GB71 tok/s86K ctx
dense
👁 Google
Gemma 2 9B
C55
9B12.7 GB32 tok/s8K ctx
dense
dolphin v2 8b abliterated i1
C55
8B7.9 GB71 tok/s86K ctx
dense
EXAONE 3.5 7.8B Instruct
C55
7.8B7.8 GB73 tok/s90K ctx
dense
aya expanse 8b orthogonal heretic i1
C55
8B7.9 GB71 tok/s86K ctx
dense
EXAONE 3.5 7.8B Instruct
C55
7.8B7.8 GB73 tok/s90K ctx
dense
HelpingAI2.5 10B i1
C55
10B9.4 GB57 tok/s52K ctx
dense
EXAONE 3.5 7.8B Instruct i1
C55
7.8B7.8 GB73 tok/s90K ctx
dense
aya expanse 8b orthogonal heretic
C55
8B7.9 GB71 tok/s86K ctx
dense
exaone 3.0 7.8b it
C55
7.8B7.8 GB73 tok/s90K ctx
dense
internlm2 5 1 8b chat i1
C55
8B7.9 GB71 tok/s86K ctx
dense
internlm3 8b instruct abliterated i1
C55
8B7.9 GB71 tok/s86K ctx
dense
Llama 2 7B Chat
C55
7B7.2 GB81 tok/s110K ctx
dense
Mistral 7B Instruct v0.2
C55
7B7.2 GB81 tok/s110K ctx
dense
Mistral 7B Instruct v0.3
C55
7B7.2 GB81 tok/s110K ctx
dense
falcon mamba 7b instruct Q4 K M
C55
7B7.2 GB102 tok/s110K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 14B
C54
14B13.6 GB29 tok/s7K ctx
dense
Mistral 7B Instruct v0.3
C54
7B7.2 GB81 tok/s110K ctx
dense
Mamba Codestral 7B v0.1
C54
7B7.2 GB102 tok/s110K ctx
dense
Hermes 2 Pro Mistral 7B
C54
7B7.2 GB81 tok/s110K ctx
dense
👁 DeepSeek
DeepSeek R1 1.5B
C54
1.5B3.4 GB24 tok/s33K ctx
dense
zephyr 7B beta
C54
7B7.2 GB81 tok/s110K ctx
dense
Nous Hermes 2 Mistral 7B DPO
C54
7B7.2 GB81 tok/s110K ctx
dense
Falcon H1R 7B
C54
7B7.2 GB81 tok/s110K ctx
dense
zephyr 7B alpha
C54
7B7.2 GB81 tok/s110K ctx
dense
speechless zephyr code functionary 7b
C54
7B7.2 GB81 tok/s110K ctx
dense
StarCoder2 7B
C54
7B7.2 GB81 tok/s110K ctx
dense
japanese stablelm instruct gamma 7B
C54
7B7.2 GB81 tok/s110K ctx
dense
Falcon H1 7B Instruct
C54
7B7.2 GB81 tok/s110K ctx
dense
👁 Unsloth
Falcon H1R 7B
C54
7B7.2 GB81 tok/s110K ctx
dense
stabilityai japanese stablelm base gamma 7b
C54
7B7.2 GB81 tok/s110K ctx
dense
baichuan2 7b chat
C54
7B7.2 GB81 tok/s110K ctx
dense
starcoder2 7b
C54
7B7.2 GB81 tok/s110K ctx
dense
zephyr 7b beta Mistral 7B Instruct v0.2
C54
7B7.2 GB81 tok/s110K ctx
dense
OpenChat 3.5 7B Qwen v2.0 i1
C54
7B7.2 GB81 tok/s110K ctx
dense
OpenChat 3.5 7B Starling v2.0 i1
C54
7B7.2 GB81 tok/s110K ctx
dense
internlm2 math plus 7b IMat
C54
7B7.2 GB81 tok/s110K ctx
dense
CodeNinja 1.0 OpenChat 7B i1
C54
7B7.2 GB81 tok/s110K ctx
dense
internlm2 5 7b chat i1
C54
7B7.2 GB81 tok/s110K ctx
dense
OpenSafetyLab MD Judge v0 2 internlm2 7b
C54
7B7.2 GB81 tok/s110K ctx
dense
MD Judge v0 2 internlm2 7b i1
C54
7B7.2 GB81 tok/s110K ctx
dense
jointpreferences mistral 7b sft helpful
C54
7B7.2 GB81 tok/s110K ctx
dense
zephyr 7b dpo full i1
C54
7B7.2 GB81 tok/s110K ctx
dense
DiscoPOP zephyr 7b gemma
C54
7B7.2 GB78 tok/s110K ctx
dense
blossom v3 baichuan2 7b i1
C54
7B7.2 GB81 tok/s110K ctx
dense
blossom v1 baichuan 7b i1
C54
7B7.2 GB81 tok/s110K ctx
dense
👁 TinyLlama
TinyLlama 1.1B
C54
1.1B3.1 GB18 tok/s4K ctx
dense
zephyr 7b gemma sft african ultrachat 100k
C54
7B7.2 GB78 tok/s110K ctx
dense
👁 01.AI
Yi 1.5 6B
C54
6B6.7 GB96 tok/s4K ctx
dense
Yi 1.5 6B Chat
C54
6B6.5 GB95 tok/s142K ctx
dense
stablelm 2 zephyr 1 6b
C53
6B6.5 GB95 tok/s142K ctx
dense
👁 LLaVA
LLaVA 1.5 7B
C53
7B14.2 GB43 tok/s4K ctx
dense
👁 BigCode
StarCoder2 7B
C53
7B6.9 GB89 tok/s16K ctx
dense
Yi 1.5 6B Chat
C53
6B6.5 GB95 tok/s142K ctx
dense
stablelm 2 1 6b chat imatrix
C53
6B6.5 GB95 tok/s142K ctx
dense
👁 NVIDIA
Nemotron Mini 4B
C53
4B6.5 GB64 tok/s4K ctx
dense
HelpingAI2 6B
C53
6B6.5 GB95 tok/s142K ctx
dense
HelpingAI2 6B i1
C53
6B6.5 GB95 tok/s142K ctx
dense
SOLAR 10.7B Instruct v1.0 uncensored
C52
10.7B9.9 GB61 tok/s43K ctx
dense
Nous Hermes 2 SOLAR 10.7B
C52
10.7B9.9 GB61 tok/s43K ctx
dense
gemma 3 12b it
C52
12B10.8 GB54 tok/s29K ctx
dense
solar finalised finetuned Model 10.7B i1
C52
10.7B9.9 GB61 tok/s43K ctx
dense
SOLAR 10.7B v1.0
C52
10.7B9.9 GB61 tok/s43K ctx
dense
👁 Google
Gemma 2 2B
C52
2B4.9 GB32 tok/s8K ctx
dense
👁 Alibaba
Qwen 2.5 1.5B
C52
1.5B3.4 GB24 tok/s131K ctx
dense
👁 MosaicML
MPT-7B-Instruct
C52
7B14.2 GB43 tok/s8K ctx
dense
👁 Baichuan
Baichuan 7B
C52
7B14.2 GB43 tok/s8K ctx
dense
HelpingAI2.5 5B i1
C51
5B5.7 GB80 tok/s187K ctx
dense
👁 Microsoft
Phi 3 Medium 14B
C51
14B13.7 GB28 tok/s7K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Qwen 14B
C51
14B12.3 GB33 tok/s13K ctx
dense
NousResearch Hermes 4 14B
C50
14B12.3 GB33 tok/s13K ctx
dense
GGUF SOLARized GraniStral 14B 2102 YeAM HCT 32QKV
C50
14B12.3 GB33 tok/s13K ctx
dense
internlm JanusCoder 14B
C50
14B12.3 GB33 tok/s13K ctx
dense
GGUF SOLARized GraniStral 14B 1902 YeAM HCT
C50
14B12.3 GB33 tok/s13K ctx
dense
👁 Unsloth
Qwen3.5 4B
C50
4B5.0 GB64 tok/s255K ctx
dense
gemma 3 4b it
C50
4B5.0 GB64 tok/s255K ctx
dense
👁 Google
Gemma 3 1B
C50
1B3.1 GB16 tok/s33K ctx
dense
gemma 3 4b it
C50
4B5.0 GB64 tok/s255K ctx
dense
Llama 3.2 3B Instruct
C48
3B4.6 GB48 tok/s352K ctx
dense
👁 Qwen
Qwen2.5 3B Instruct
C48
3B4.3 GB48 tok/s367K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 0.5B
C48
0.5B2.6 GB8 tok/s131K ctx
dense
Llama 3.2 3B Instruct
C48
3B4.3 GB48 tok/s367K ctx
dense
Ministral 3 3B Instruct 2512
C48
3B4.3 GB48 tok/s367K ctx
dense
HELVETE 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
Hermes 3 Llama 3.2 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
stablelm zephyr 3b
C47
3B4.3 GB48 tok/s367K ctx
dense
StarCoder2 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
AI21 Jamba Reasoning 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
stablelm 3b 4e1t
C47
3B4.3 GB48 tok/s367K ctx
dense
ai21labs AI21 Jamba Reasoning 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
ai21labs AI21 Jamba2 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
HelpingAI 3B hindi i1
C47
3B4.3 GB48 tok/s367K ctx
dense
AI21 Jamba2 3B
C47
3B4.3 GB48 tok/s367K ctx
dense
HelpingAI 3B hindi
C47
3B4.3 GB48 tok/s367K ctx
dense
AI21 Jamba2 3B i1
C47
3B4.3 GB48 tok/s367K ctx
dense
👁 Alibaba
Qwen 3 0.6B
C47
0.6B3.3 GB10 tok/s33K ctx
dense
👁 Meta
Llama 3.2 1B
C47
1B3.2 GB16 tok/s128K ctx
dense
👁 BigCode
StarCoder2 3B
C46
3B4.4 GB48 tok/s16K ctx
dense
gemma 2 2b it
C46
2B4.0 GB32 tok/s564K ctx
dense
EXAONE 3.5 2.4B Instruct
C46
2.4B3.8 GB38 tok/s480K ctx
dense
👁 Google
gemma 2b
C45
2B3.6 GB32 tok/s593K ctx
dense
Gemmasutra Mini 2B v1
C45
2B3.6 GB32 tok/s593K ctx
dense
gemma 2 2b it
C45
2B3.6 GB32 tok/s593K ctx
dense
👁 Alibaba
Qwen 3.5 0.6B
C44
0.6B3.3 GB10 tok/s131K ctx
dense
👁 Qwen
Qwen2.5 1.5B Instruct
C44
1.5B3.2 GB24 tok/s818K ctx
dense
👁 Unsloth
DeepSeek R1 Distill Qwen 1.5B
C43
1.5B3.2 GB24 tok/s818K ctx
dense
Yi Coder 1.5B Chat
C43
1.5B3.2 GB24 tok/s818K ctx
dense
stablelm 2 zephyr 1.6b
C43
1.6B3.3 GB26 tok/s762K ctx
dense
logos16v2 stablelm2 1.6b i1
C43
1.6B3.3 GB26 tok/s762K ctx
dense
Yi Coder 1.5B
C43
1.5B3.2 GB24 tok/s818K ctx
dense
👁 Unsloth
Falcon H1 1.5B Instruct
C43
1.5B3.2 GB24 tok/s818K ctx
dense
EXAONE 4.0 1.2B
C42
1.2B3.0 GB19 tok/s1043K ctx
dense
TinyLlama 1.1B Chat v1.0
C42
1.1B2.9 GB18 tok/s1146K ctx
dense
Llama 3.2 1B Instruct Q8 0
C42
1B3.0 GB16 tok/s1240K ctx
dense
TinyLlama 1.1B Chat v0.3
C42
1.1B2.9 GB18 tok/s1146K ctx
dense
👁 TinyLlama
TinyLlama 1.1B Chat v0.6
C42
1.1B2.9 GB18 tok/s1146K ctx
dense
👁 Alibaba
Qwen 2.5 0.5B
C41
0.5B2.6 GB8 tok/s131K ctx
dense
Llama 3.2 1B Instruct
C41
1B2.8 GB16 tok/s1268K ctx
dense
gemma 3 1b it
C41
1B2.8 GB16 tok/s1268K ctx
dense
TinyLlama 1.1B Chat v1.0 imatrix
C41
1.1B2.9 GB18 tok/s1146K ctx
dense
Falcon3 1B Instruct abliterated
C41
1B2.8 GB16 tok/s1268K ctx
dense
👁 DeepSeek
DeepSeek LLM 7B
C40
7B13.7 GB46 tok/s4K ctx
dense
F

F Tier -- Can't Run (142 models)

Too heavy for this hardware.

Phi 4 reasoning vision 15B
D40
15B13.0 GB27 tok/s7K ctx
dense
StarCoder2 15B
D40
15B13.0 GB27 tok/s7K ctx
dense
starcoder2 15b instruct v0.1
D40
15B13.0 GB27 tok/s7K ctx
dense
starcoder2 15b i1
D40
15B13.0 GB27 tok/s7K ctx
dense
starcoder2 15b instruct v0.1
D40
15B13.0 GB27 tok/s7K ctx
dense
HelpingAI 15B i1
D39
15B13.0 GB27 tok/s7K ctx
dense
SmolVLM 500M Instruct
D39
0.5B2.6 GB8 tok/s1518K ctx
dense
embeddinggemma 300M
D37
0.3B2.4 GB5 tok/s1545K ctx
dense
👁 LMSYS
Vicuna 7B
D36
7B14.2 GB43 tok/s4K ctx
dense
👁 BigCode
StarCoder2 15B
D35
15B14.1 GB21 tok/s4K ctx
dense
granite embedding 107m multilingual
D34
0.11B2.3 GB2 tok/s1574K ctx
dense
Falcon H1 Tiny 90M Instruct
D34
0.09B2.3 GB2 tok/s1575K ctx
dense
👁 Meta
Llama 3.1 70B
F0
70B49.7 GB2 tok/s4K ctx
dense
👁 Meta
Llama 3.3 70B
F0
70B49.7 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 32B
F0
32B25.5 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 72B
F0
72B50.9 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 Coder 32B
F0
32B25.5 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 3 32B
F0
32B25.5 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 3 235B A22B
F0
235B148.3 GB2 tok/s4K ctx
moe
👁 Alibaba
Qwen 3 30B A3B
F0
30.5B22.2 GB9 tok/s4K ctx
moe
👁 Alibaba
Qwen3-Coder 30B A3B Instruct
F0
30.5B22.2 GB9 tok/s4K ctx
moe
👁 Alibaba
Qwen3-Coder-Next
F0
80B52.4 GB4 tok/s4K ctx
moe
👁 Alibaba
Qwen3-Coder 480B A35B Instruct
F0
480B297.8 GB2 tok/s4K ctx
moe
👁 Alibaba
Qwen3-VL 30B A3B Instruct
F0
30B21.9 GB13 tok/s4K ctx
moe
👁 Alibaba
Qwen 3.5 27B
F0
27B21.7 GB4 tok/s4K ctx
dense
👁 Alibaba
Qwen 3.5 35B A3B
F0
35B24.9 GB9 tok/s4K ctx
moe
👁 Alibaba
Qwen 3.6 35B A3B
F0
35B27.6 GB8 tok/s4K ctx
+1moe
👁 Alibaba
Qwen 3.6 27B
F0
27B19.5 GB3 tok/s4K ctx
+1dense
👁 Alibaba
Qwen 3.5 122B A10B
F0
122B79.0 GB2 tok/s4K ctx
moe
👁 Alibaba
Qwen 3.5 397B A17B
F0
397B247.1 GB2 tok/s4K ctx
moe
👁 Google
Gemma 2 27B
F0
27B29.8 GB2 tok/s4K ctx
dense
👁 Google
Gemma 3 27B
F0
27B29.8 GB2 tok/s4K ctx
dense
👁 Mistral
Mistral Small 24B
F0
24B19.2 GB6 tok/s4K ctx
dense
👁 Mistral
Mistral Small 3.2 24B
F0
24B19.2 GB6 tok/s4K ctx
vision
👁 Mistral
Mixtral 8x7B
F0
47B32.7 GB4 tok/s4K ctx
moe
👁 Mistral
Mixtral 8x22B
F0
141B91.5 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek R1 Distill 32B
F0
32B25.5 GB2 tok/s4K ctx
dense
👁 IBM
Granite 4.1 30B
F0
30B24.3 GB2 tok/s4K ctx
dense
👁 Cohere
Command R 35B
F0
35B25.9 GB2 tok/s4K ctx
dense
👁 TII
Falcon 40B Instruct
F0
40B32.7 GB2 tok/s4K ctx
dense
👁 WizardLM
WizardLM 13B
F0
13B22.2 GB10 tok/s4K ctx
dense
👁 BigCode
StarCoder 15B
F0
15B27.5 GB6 tok/s4K ctx
dense
👁 Meta
CodeLlama 13B Instruct
F0
13B22.2 GB10 tok/s4K ctx
dense
👁 MosaicML
MPT-30B-Instruct
F0
30B47.1 GB2 tok/s4K ctx
dense
👁 Mistral
Devstral Small 1.1
F0
24B19.2 GB6 tok/s4K ctx
dense
👁 Mistral
Devstral Small 2 24B Instruct
F0
24B19.2 GB6 tok/s4K ctx
dense
👁 Mistral
Devstral 2 123B Instruct
F0
123B82.5 GB2 tok/s4K ctx
dense
👁 Mistral
Mistral Small 4 119B
F0
119B80.1 GB3 tok/s4K ctx
moe
👁 Mistral
Leanstral 119B A6B
F0
119B83.5 GB2 tok/s4K ctx
moe
👁 Mistral
Mistral Large 3
F0
675B419.2 GB2 tok/s4K ctx
+1moe
👁 Mistral
Magistral Small 2507
F0
24B19.2 GB6 tok/s4K ctx
dense
👁 Mistral
Codestral 2 25.08
F0
22B18.0 GB7 tok/s4K ctx
dense
F0
744B475.0 GB2 tok/s4K ctx
moe
👁 Z.ai
GLM-5.1
F0
754B481.1 GB2 tok/s4K ctx
moe
F0
1000B619.5 GB2 tok/s4K ctx
moe
F0
1000B619.5 GB2 tok/s4K ctx
+1moe
👁 Moonshot AI
Kimi Linear 48B A3B
F0
48B32.3 GB2 tok/s4K ctx
linear
👁 Nous Research
Nous Hermes 1.0
F0
9B19.8 GB17 tok/s6K ctx
dense
👁 Nous Research
Nous Dolphin 13B
F0
13B23.7 GB8 tok/s4K ctx
dense
👁 Baichuan
Baichuan 13B
F0
13B23.7 GB8 tok/s4K ctx
dense
👁 Cerebras
Cerebras-GPT 13B
F0
13B21.2 GB10 tok/s4K ctx
dense
👁 Stability AI
StableLM 2 12B
F0
12B22.9 GB8 tok/s4K ctx
dense
👁 InternLM
InternLM 20B
F0
20B34.8 GB3 tok/s4K ctx
dense
👁 01.AI
Yi 34B Chat
F0
34B26.5 GB2 tok/s4K ctx
dense
👁 DeepSeek
DeepSeek V3 671B
F0
671B471.0 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V3.1 671B
F0
671B471.0 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V4 Pro
F0
1600B866.0 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V4 Flash
F0
284B161.4 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek R1 671B
F0
671B471.0 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek Coder V2 16B
F0
16B15.2 GB38 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek Coder V2 236B
F0
236B204.7 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek V2.5 236B
F0
236B204.7 GB2 tok/s4K ctx
moe
👁 DeepSeek
DeepSeek LLM 67B
F0
67B48.8 GB2 tok/s4K ctx
dense
👁 Meta
Llama 4 Scout 17B 16E
F0
109B71.5 GB2 tok/s4K ctx
moe
👁 Meta
Llama 4 Maverick 17B 128E
F0
400B249.0 GB2 tok/s4K ctx
moe
👁 Cohere
Command R+ 104B
F0
104B69.0 GB2 tok/s4K ctx
dense
👁 Cohere
Command A 111B
F0
111B73.7 GB2 tok/s4K ctx
dense
👁 01.AI
Yi 1.5 34B
F0
34B26.5 GB2 tok/s4K ctx
dense
👁 Mistral AI
Codestral 22B
F0
22B18.0 GB7 tok/s4K ctx
dense
👁 Mistral AI
Pixtral Large 124B
F0
124B83.1 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 VL 72B
F0
72B50.9 GB2 tok/s4K ctx
dense
👁 Alibaba
Qwen 2.5 Math 72B
F0
72B50.9 GB2 tok/s4K ctx
dense
👁 NVIDIA
Nemotron 70B
F0
70B49.7 GB2 tok/s4K ctx
dense
👁 LLaVA
LLaVA 1.6 13B
F0
13B22.2 GB10 tok/s4K ctx
dense
F0
19B16.1 GB12 tok/s4K ctx
dense
👁 LMSYS
Vicuna 13B
F0
13B22.2 GB10 tok/s4K ctx
dense
👁 IBM
Granite Code 20B
F0
20B17.5 GB9 tok/s4K ctx
dense
👁 OpenAI
GPT-OSS 20B
F0
21B17.4 GB23 tok/s4K ctx
moe
👁 Allen AI
OLMo 2 32B
F0
32B25.5 GB2 tok/s4K ctx
dense
👁 NVIDIA
Nemotron 3 Nano 30B
F0
30B22.8 GB3 tok/s4K ctx
dense
👁 LG AI
EXAONE 4.0 32B
F0
32B25.5 GB2 tok/s4K ctx
dense
F0
24B19.2 GB6 tok/s4K ctx
dense
👁 Meta
Llama 3.1 405B
F0
405B256.8 GB2 tok/s4K ctx
dense
👁 DeepSeek
DeepSeek R1 Distill 70B
F0
70B49.7 GB2 tok/s4K ctx
dense
👁 IBM
Granite Code 34B
F0
34B26.5 GB2 tok/s4K ctx
dense
👁 OpenAI
GPT-OSS 120B
F0
117B78.4 GB2 tok/s4K ctx
dense
👁 Cohere
Aya Expanse 32B
F0
32B24.1 GB2 tok/s4K ctx
dense
👁 Mistral AI
Mistral Small 3.1 24B
F0
24B19.2 GB6 tok/s4K ctx
dense
👁 Google
Gemma 4 26B A4B
F0
25.2B21.1 GB13 tok/s4K ctx
moe
👁 Google
Gemma 4 31B
F0
30.7B35.5 GB2 tok/s4K ctx
dense
👁 DeepSeek
DeepSeek V3.2
F0
671B411.9 GB2 tok/s4K ctx
moe
👁 NVIDIA
Nemotron Cascade 2 30B A3B
F0
30B23.3 GB11 tok/s4K ctx
moe
MiniMax M2.7
F0
230B146.2 GB2 tok/s4K ctx
moe
Solar Open 100B
F0
100B74.8 GB2 tok/s4K ctx
dense
Baichuan M2 32B Q4 K M
F0
32B25.4 GB2 tok/s4K ctx
dense
baichuan inc Baichuan M2 32B
F0
32B25.4 GB2 tok/s4K ctx
dense
Codestral 22B v0.1
F0
22B18.1 GB7 tok/s4K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
F0
24B19.6 GB5 tok/s4K ctx
dense
cognitivecomputations Dolphin3.0 R1 Mistral 24B
F0
24B19.6 GB5 tok/s4K ctx
dense
internlm2 5 20b chat
F0
20B16.6 GB10 tok/s4K ctx
dense
Qwen3 48B A4B Savant Commander Distill 12X Closed Open Heretic Uncensored
F0
48B37.0 GB2 tok/s4K ctx
dense
internlm2 limarp chat 20b
F0
20B16.6 GB10 tok/s4K ctx
dense
Codestral 22B v0.1 IMat
F0
22B18.1 GB7 tok/s4K ctx
dense
EXAONE 4.0 32B
F0
32B25.4 GB2 tok/s4K ctx
dense
K EXAONE 236B A23B
F0
236B173.7 GB2 tok/s4K ctx
dense
Codestral 22B v0.1
F0
22B18.1 GB7 tok/s4K ctx
dense
Qwen3.5 35B A3B
F0
35B27.6 GB2 tok/s4K ctx
dense
gemma 3 27b it
F0
27B21.7 GB3 tok/s4K ctx
dense
Llama 3.3 70B Instruct
F0
70B53.0 GB2 tok/s4K ctx
dense
Mistral Small 24B Instruct 2501
F0
24B19.6 GB5 tok/s4K ctx
dense
mistral small 3.1 24b instruct 2503 hf
F0
24B19.6 GB5 tok/s4K ctx
dense
aya expanse 32b heretic MPOA i1
F0
32B25.4 GB2 tok/s4K ctx
dense
Baichuan M3 235B
F0
235B173.0 GB2 tok/s4K ctx
dense
Baichuan M3 235B i1
F0
235B173.0 GB2 tok/s4K ctx
dense
BaichuanMed OCR 72B i1
F0
72B54.5 GB2 tok/s4K ctx
dense
Codestral 21B Pruned i1
F0
21B17.4 GB8 tok/s4K ctx
dense
Codestral 22B v0.1 i1
F0
22B18.1 GB7 tok/s4K ctx
dense
Codestral RAG 19B Pruned i1
F0
19B15.9 GB11 tok/s4K ctx
dense
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored i1
F0
24B19.6 GB5 tok/s4K ctx
dense
internlm2 math plus 20b i1
F0
20B16.6 GB10 tok/s4K ctx
dense
Solar Open 100B i1
F0
100B74.8 GB2 tok/s4K ctx
dense
Solar Open 69B REAP i1
F0
69B52.3 GB2 tok/s4K ctx
dense
Hermes 4.3 36B
F0
36B28.3 GB2 tok/s4K ctx
dense
stabilityai japanese stablelm instruct beta 70b
F0
70B53.0 GB2 tok/s4K ctx
dense
Codestral 22B v0.1
F0
22B18.1 GB7 tok/s4K ctx
dense
👁 Unsloth
gemma 3 27b it
F0
27B21.7 GB3 tok/s4K ctx
dense
👁 Unsloth
Mistral Small 3.2 24B Instruct 2506
F0
24B19.6 GB5 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 122B A10B
F0
122B76.2 GB2 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 27B
F0
27B21.7 GB3 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 35B A3B
F0
35B27.6 GB2 tok/s4K ctx
dense
👁 Unsloth
Qwen3.5 397B A17B
F0
397B290.8 GB2 tok/s4K ctx
dense
cognitivecomputations Dolphin Mistral 24B Venice Edition
F0
24B19.6 GB5 tok/s4K ctx
dense

Compare with...

See How Other Hardware Stacks Up

Check tier lists for similar GPUs and Apple Silicon configs

RTX 3060 12GB

12 GB VRAM

View tier list

RTX 3080 Ti 12GB

12 GB VRAM

View tier list

RTX 4070 Super 12GB

12 GB VRAM

View tier list

RTX 5070 12GB

12 GB VRAM

View tier list
Full RTX 4070 12GB pageCompare this hardware